|Publication number||US7720572 B2|
|Application number||US 11/541,386|
|Publication date||May 18, 2010|
|Filing date||Sep 29, 2006|
|Priority date||Sep 30, 2005|
|Also published as||EP1941411A2, EP1941411B1, EP2050544A1, EP2050544B1, EP2281667A1, EP2281667B1, EP2281668A1, EP2281668B1, US7957837, US8195333, US8583282, US8935006, US9446510, US9452525, US9796078, US20070192910, US20070198128, US20070199108, US20090177323, US20110172822, US20120303160, US20140039680, US20140095007, US20140142757, US20150224640, WO2007041295A2, WO2007041295A3, WO2007041295A8|
|Publication number||11541386, 541386, US 7720572 B2, US 7720572B2, US-B2-7720572, US7720572 B2, US7720572B2|
|Inventors||Andrew Ziegler, Andrew Jones, Clara Vu, Matthew Cross, Ken Sinclair, Tony L. Campbell|
|Original Assignee||Irobot Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (13), Non-Patent Citations (5), Referenced by (89), Classifications (47), Legal Events (5)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This application claims priority to and incorporates by reference herein in their entireties, the following: U.S. Provisional Patent Application Ser. No. 60/746,491, filed May 4, 2006; U.S. Provisional Patent Application Ser. No. 60/745,006, filed Apr. 17, 2006; U.S. Provisional Patent Application Ser. No. 60/722,935, filed Sep. 30, 2005.
The present invention relates generally to autonomous mobile robots for interacting with people and, more specifically, to autonomous mobile robots for assisting persons with various tasks.
Robotic research platforms have been developed for interacting with people in home situations, such as the elderly, children, or others who may benefit from an interactive robot assistant. These robotic platforms often do not consider the actual home environment, or personal preferences and concerns, such as making a companion robot non-intrusive and a welcome guest in the home environment.
Certain platforms have been developed which assist caregivers in carrying medications, providing amusing interaction, and/or providing teleconferencing tools, many remain research platforms: they tend to be so large, heavy, and unwieldy, as to be inappropriate for use in an ordinary home. These robotic platforms are generally tested in institutions, where some surveillance is expected and privacy and personal dignity tend already to be adversely affected, and may not be suitable for use in private homes, where expectations for privacy protection are higher.
Simply making a robot smaller, however, does not correct these shortcomings. Though a robot may be sufficiently small to be carried, the contemplated use is typically in a single room, with limited ability to be useful throughout an entire home. Low-weight platforms tend also to be low to the ground, and generally out of view of residents who may be moving about in the same rooms, creating a danger for those with balance or mobility issues. In short, previous robots suffer from limited usability, and thus do not function as a welcoming and versatile assistant to a human.
In one aspect, the invention relates to a robot for regimen compliance management, including a processor, a memory accessible by the processor, a sensor capable of detecting a presence of a person within a detection range of the robot, a communication interface capable of generating a human-perceptible signal, a drive operatively connected to the processor that moves the robot, a scheduler routine executable on the processor that checks at least one of a medication dosage information for a medication event and a health-related information for a regimen event, and in advance of the event, initiates a person finding routine, a person finding routine executable on the processor that instructs the drive to move the robot about an environment and to stop in a position proximate a person, and a regimen compliance manager for ensuring compliance of a person with a regimen routine.
In one aspect, the invention relates to a robot for regimen compliance management, including a processor, a memory accessible by the processor that includes personal medication dosage information, a sensor capable of detecting the presence of a person within a detection range of the robot, a communication interface capable of generating a human-perceptible signal and receiving a compliance or non-compliance indication from the person, a drive operatively connected to the processor that moves the robot, a scheduler routine executable on the processor that checks the personal medication do sage information for medication events, and in advance of a medication event, initiates a person finding routine, a person finding routine executable on the processor that instructs the drive to move the robot about the household checking the sensor suite for the presence of the person and to stop in a position next to the person, a matching routine executable on the processor that sends a human perceptible signal to the person including information regarding the location of medication to be taken and guides the person to the medication. In certain embodiments of the above aspect, the communication interface is receiving a compliance or non-compliance indication from the person and the memory accessible by the processor includes postponement rules, and further including a regimen compliance snooze routine that upon recognition of a non-compliance indication from the person, checks the personal medication dosage information and sets a new medication event compatible with the postponement rules.
In another aspect, the invention relates to a robot for regimen compliance management, including a processor, a memory accessible by the processor that includes personal medication dosage information, the memory including postponement rules defining permissible conditions for postponing dosage, a sensor capable of detecting the presence of a person within a detection range of the robot, a communication interface capable of generating a human-perceptible signal and receiving a compliance or non-compliance indication from the person, a drive operatively connected to the processor that moves the robot, a scheduler routine executable on the processor that checks the personal medication dosage information for medication events, and in advance of a medication event, initiates a person finding routine, a person finding routine executable on the processor that instructs the drive to move the robot about the household and to stop in a position next to the person, a matching routine executable on the processor that sends a human perceptible signal to the person including information regarding medication to be taken, and a regimen compliance snooze routine that upon recognition of a non-compliance indication from the person, checks the personal medication dosage information and sets a new medication event if postponing medication is compatible with the postponement rules.
In yet another aspect, the invention relates to a robot for regimen compliance management, including a medication cache that receives loaded medication to be carried by the robot, a processor, a memory accessible by the processor that includes personal medication dosage information and postponing rules, a sensor capable of detecting the presence of a person within a detection range of the robot, a communication interface capable of generating a human-perceptible signal, a drive operatively connected to the processor that moves the robot, a scheduler routine executable on the processor that checks the personal medication dosage information for medication events, and in advance of a medication event, initiates a person finding routine, a person finding routine executable on the processor that instructs the drive to move the robot about the household checking the sensor suite for the presence of the person and to stop in a position next to the person, a matching routine executable on the processor that sends a human perceptible signal to the person including information regarding the medication carried in the medication cache. In certain embodiments of the above aspect, the robot includes a medication loading routine executable on the processor that sends a human perceptible signal to the person including information regarding loading the medication cache for later administration and information guiding the person to the medication.
In still another aspect, the invention relates to a robot system for regimen compliance management, having a medication cache that receives loaded medication, a processor, a memory accessible by the processor that includes personal medication dosage information and postponing rules, a sensor capable of detecting the presence of a person within a detection range of the robot, a communication interface capable of generating a human-perceptible signal, a drive operatively connected to the processor that moves the robot, a scheduler routine executable on the processor that checks the personal medication dosage information for medication events, and in advance of a medication event, initiates a person finding routine, a person finding routine executable on the processor that instructs the drive to move the robot about the household checking the sensor suite for the presence of the person and to stop in a position next to the person, and a medication loading routine executable on the processor that sends a human perceptible signal to the person including information regarding loading the medication cache for later administration, and information guiding the person to the medication cache.
In another aspect, the invention relates to a robot for regimen compliance management, including a network interface that connects the robot to a remote location at which a caregiver may connect to the robot, a processor, a memory accessible by the processor that includes personal medication dosage information, a sensor capable of detecting the presence of a person within a detection range of the robot, a communication interface capable of generating a human-perceptible signal and receiving a compliance or non-compliance indication from the person, a drive operatively connected to the processor that moves the robot, a scheduler routine executable on the processor that checks the personal medication dosage information for medication events, and in advance of a medication event, initiates a person finding routine, a person finding routine executable on the processor that instructs the drive to move the robot about the household checking the sensor suite for the presence of the person and to stop in a position next to the person, a regimen compliance reminder routine executable on the processor that, upon recognition of a non-compliance indication from the person, contacts the caregiver via the network interface. In certain embodiments of the above aspect, the robot includes an inbound communication channel for the caregiver to send a human-perceptible signal through the inbound communication channel and via the communication interface.
In yet another aspect, the invention relates to a robot for regimen compliance management, including a network interface that connects the robot to a remote location at which a caregiver may connect to the robot, a processor, a memory accessible by the processor that includes health-related regimen information, a sensor capable of detecting the presence of a person within a detection range of the robot, a communication interface capable of generating a human-perceptible signal, an inbound communication channel for the caregiver to send a human-perceptible signal through the inbound communication channel and via the communication interface, a drive operatively connected to the processor that moves the robot, a scheduler routine executable on the processor that checks the health-related regimen information for regimen events, and in advance of a health-related regimen event, initiates a person finding routine, a person finding routine executable on the processor that instructs the drive to move the robot about the household checking the sensor suite for the presence of the person and to stop in a position next to the person, and a regimen compliance guide access routine executable on the processor that connects a communication session with a caregiver via the network interface.
In another aspect, the invention relates to a method of human-robot interaction, including receiving a communication script segment, outputting to a person at least one of a visible component of the communication script segment and an audible component of the communication script segment, and controlling at least one of a robot expression component accompanying the output component and a robot response to an input by a person.
In still another aspect, the invention relates to a method of human-robot interaction, including receiving a communication script segment, including an output query sub-script text and a response tree of five or less sub-script response text candidate, associating the output query text with an audible output signal, and outputting the audible output signal as a spoken query to a person, displaying the output query sub-script text together with the five or less sub-script response text candidates on a display of the robot, receiving an audio input signal recording a person's response to the audible output signal, processing the audio input signal to recognize if the audio input signal includes speech corresponding to any one of the five or less displayed sub-script response text candidates, if the audio input signal is not recognized to include speech corresponding to any one of the four or less sub-script response text candidates, issuing an output signal to prompt the user to retry communicating a response to the audible output signal, and if the audio input signal is recognized to include speech corresponding to any one of the five or less sub-script response text candidates, issuing an output signal including a repetition of the one of the five or less sub-script response text candidates that was recognized. Embodiments of the above aspect include, after the confirmation signal is issued monitoring for a confirmation of the correctness or incorrectness of the one of the four or less sub-script response text candidates that was recognized. In still other embodiments issuing an output signal to prompt the user to retry communicating a response to the audio output signal includes, highlighting the displayed sub-script text on a display of the robot, and receiving an input signal selecting any one of the five or less displayed sub-script response text candidates via a manually operated control associated with the display. In additional embodiments two of the five or less sub-scripts response text candidates are a simple affirmative and simple negative response, and further include processing the audio input signal to recognize if the audio input signal includes speech corresponding to a family of affirmative responses equivalent to the simple affirmative response or includes speech corresponding to a family of affirmative responses equivalent to the simple negative response. In other embodiments one of the five or less sub-scripts response text candidates is a response text candidate that is common to a majority of communication script segments. In still other embodiments, no two of the five or less sub-scripts response text candidates are a simple affirmative and simple negative response, and further include processing the audio input signal to recognize if the audio input signal includes speech corresponding to a family of affirmative responses equivalent to the simple affirmative response or includes speech corresponding to a family of affirmative responses equivalent to the simple negative response. In other embodiments of the robot, the response tree is a response tree of three or less sub-script response text candidates, and none of the response text candidates is simple affirmative response or simple negative response. Other embodiments include converting the response tree text subscript into a set of menu choices displayed on the display.
In another aspect, the invention relates to a method of human-robot interaction, including receiving a communication script segment, including an output query sub-script text and a response tree of five or less sub-script response text candidate, associating the output query text with an audible output signal, and outputting the audible output signal as a spoken query to a person, displaying the output query sub-script text together with the five or less sub-script response text candidates on a display of the robot, receiving an audio input signal recording a person's response to the audible output signal, processing the audio input signal to recognize if the audio input signal includes speech corresponding to any one of the five or less displayed sub-script response text candidates, if the audio input signal is recognized to include speech corresponding to any one of the five or less sub-script response text candidates, issuing an output signal including a repetition of the one of the five or less sub-script response text candidates that was recognized if the audio input signal is not recognized to include speech corresponding to any one of the four or less sub-script response text candidates, highlighting the displayed sub-script text on a display of the robot, then receiving an input signal selecting any one of the five or less displayed sub-script response text candidates via a manually operated control associated with the display.
In yet another aspect, the invention relates to a method of human-robot interaction, including asynchronously executing a plurality of motion behaviors, including motion behaviors responsive to events, receiving a communication script segment, including dialogue branches with robot speech prompt text and human response text, interpreting the communication script segment to generate an audible robot speech prompt, receiving input from a person as a response to the audible robot speech prompt, interrupting a dialogue branch in response to an event detected by one of the plurality of behaviors to execute an asynchronous response, and recovering the dialogue branch after the execution of the asynchronous response.
In still another aspect, the invention relates to a method of human-robot interaction, including receiving a communication script segment, including dialogue branches with robot speech prompt text and human response text, associating the output query text with an audible output signal, and outputting the audible output signal as a spoken query to a person, modulating the dialogue branch to show a desired expression during the communication script segment by adding at least expression motion selected from a head movement sequence including nod axis head movement or turn axis head movement, or a robot movement sequence including movement of the entire robot.
In another aspect, the invention relates to a method of human-robot interaction, including receiving a communication script segment, including dialogue branches with robot speech prompt text, human response text, and robot expression motion tags, interpreting the communication script segment to generate an audible robot speech prompt and robot expression motions according to the dialogue branches, receiving input from a person as a response to the audible robot speech prompt, interrupting an expression motion in response to an event detected by the robot to execute a corrective behavior to reposition the robot according to the event, and recovering the dialogue branch after the execution of the corrective behavior.
In another aspect, the invention relates to a method of robot self-navigation for a robot, the method including monitoring at least one of a sensor and an input, comparing a signal from at least one of the sensor and the input to at least two predetermined conditions, and performing at least one of (a) an action, if the signal corresponds to a first condition; and (b) a movement of the robot, if the signal corresponds to a second condition.
In one aspect, the invention relates to a method of robot self-navigation, including monitoring a sensor capable of detecting the presence of a person within a detection range of the robot, improving a presence score when the person is detected within the detection range of the robot, decaying the presence score to progressively worsen, driving in a direction to move the robot to a different location when the presence score decays to a first threshold presence score, parking the robot in a location proximate to the person when the presence score improves to a second threshold presence score.
In another aspect, the invention relates to a method of robot self-navigation, including monitoring a sensor capable of detecting the presence of a detectable object within a detection range of the robot, improving a presence score when a detectable object is detected within the detection range of the robot, decaying the presence score to progressively worsen, driving in a direction to move the robot to a different location when the presence score decays to a first threshold presence score. In certain embodiments, the sensor suite is capable of detecting detectable objects including a person and a charging station, and the method further includes parking the robot in a charging station located proximate to the person, engaged to recharge, when the presence score improves to a second threshold presence score. In other embodiments, the driving includes moving the robot to a succession of different locations until the presence score improves to a second threshold presence score.
In yet another aspect, the invention relates to a method of robot self-navigation, including monitoring a sensor suite capable of detecting the presence of a detectable object within a detection range of the robot, detectable objects including either or both of a person and a charging station, improving a presence score when a detectable object is detected within the detection range of the robot, decaying the presence score to progressively worsen, driving in a direction to move the robot to a different location when the presence score decays to a first threshold presence score, and parking the robot in a charging station located proximate to the different location and to the person, engaged to recharge, when the presence score improves to a second threshold presence score.
In another aspect, the invention relates to a mobile robot for interacting with a person, including a first detector for detecting at least one of a person and an object within a first predetermined space proximate the robot, a second detector for detecting at least one of a person and an object within a second predetermined space proximate the robot, a physical contact device to detect contact between at least one of a person and an object, wherein at least one of the first detector, the second detector, and the contact device triggers a safety condition, and a controller configured to immobilize the robot when the safety condition is detected.
In still another aspect, the invention relates to a method of robot self-navigation, including generating an annunciator activation signal in response to a detected condition, activating an annunciator upon an annunciator activation signal, monitoring a sensor for a detected response from a person, checking whether the detected response includes a sought response, if the checking confirms the detected response includes a sought response, activating a found person success routine, and if the checking confirms the detected response includes a sought response, identifying a passage from a present chamber, and driving in a direction to move the robot through the passage from the present chamber to a new chamber.
It is a preferably a function of the robot to act as a confidence, trust, and bonding manager. The robot is crafted to boost user trust and emotional attachment, which is beneficial in itself and provides generic benefits for any and all other applications, increasing their success rate. This function addresses the challenge of introducing new and unfamiliar technology (especially autonomous, self-controlled, mobile technology) into the lives of ordinary persons.
In another aspect, the invention relates to a method for robotically locating a person being sought, including monitoring a sensor suite for a non-noise signal indicative of the presence of a person, and if the non-noise signal corresponds to a sought response, then activating a found person success routine, and if the non-noise signal does not correspond to a sought response or if no response is identified, then selecting one of (i) identifying a doorway from a present chamber, and navigating the robot through the doorway from the present chamber to a new chamber, or (ii) controlling a mobile robot to activate an annunciator. In certain embodiments of the above-aspect, the annunciator emits a sound including the name of the person being sought, and the method further includes checking the sought response to confirm that the source of the sought response is a person of the name being sought.
In yet another aspect, the invention relates to a method for navigating a robot, including receiving a first navigation command representative of a remote user's selection of a room identity marker, recognizing a present room, driving among rooms of different room identity according to the first navigation command until the robot recognizes being within a room having a room identity corresponding to the selected room identity marker, receiving a second navigation command representative of a remote user's selection of one of a floor location within the room having a room identity corresponding to the room identity marker or a landmark item within the room having a room identity corresponding to the room identity marker, driving within the room corresponding to the room identity according to the second navigation command until the robot recognizes being at one of the floor location or next to the landmark item, receiving a third navigation command stream representative of a remote user's command including direction and bearing, driving within the room and from the floor location according to the received third navigation command stream representative of a remote user's command including direction and bearing.
In still another aspect, the invention relates to a method for handling navigation commands for remotely controlling a robot, including receiving a selection of room identities corresponding to available room identities in a household, driving among rooms of different room identity until the robot captures topological adjacency among room identities and correlates each room identity of a household with a received room identity marker, displaying room identity markers corresponding to available room identities for a household, the room identity markers being displayed according to topological adjacency of corresponding room identities and being displayed as a part of a graphical depiction of a house.
In another aspect, the invention relates to a method for handling navigation commands for remotely controlling a robot, including receiving a selection of room identities corresponding to available room identities in a household, driving among rooms of different room identity until the robot captures topological adjacency among room identities and correlates each room identity of a household with a received room identity marker, displaying room identity markers corresponding to available room identities for a household, the room identity markers being displayed according to topological adjacency of corresponding room identities and being displayed as a part of a graphical depiction of a house, receiving a selection of a room identity marker via a user interface linked to the displaying room identity markers as a first navigation command representative of a remote user's selection of a room identity marker, recognizing a present room, and driving among rooms of different room identity according to the first navigation command until the robot recognizes being within a room having a room identity corresponding to the selected room identity marker.
In yet another aspect, the invention relates to a mobile robot for interacting with a person, including an annunciator configured to audibly indicate the presence of the robot, a visible beacon configured to optically indicate the presence of the robot, a proximity curtain detector configured to detect a person within a curtain covering a majority of the height of the robot and thereafter trigger a safety condition, a physical impact buffer configured to absorb impact of a person colliding with the robot, detect the person, and trigger the safety condition, a floor-level proximity ribbon detector configured to detect objects of less than 2 inches in height and trigger the safety condition, and a controller configured to immobilize the robot when the safety condition is detected.
In still another aspect, the invention relates to a mobile robot for interacting with a person, including a zone detector configured to detect an object within a zone in front of the robot covering a substantial portion of the height of the robot above 10 inches from the ground and thereafter trigger a safety condition, a near zone impact bumper configured to absorb an impact of an object colliding with the robot, detect the object, and trigger the safety condition, a near zone proximity sensor configured to detect an object in front of the robot between 2-10 inches from the ground and trigger the safety condition, a near zone low object proximity sensor configured to detect objects in front of the robot of less than 2-4 inches in height and trigger the safety condition, and a controller configured to immobilize the robot when the safety condition is detected. In embodiments of the above aspect, each of the proximity curtain detector, the physical impact buffer, and the floor-level proximity ribbon detector are configured to detect a person at different distances from a center of the robot. In certain embodiments, the robot further includes a heat radiation detector or sound detector checked by the controller, that determines that an object encountered in the direction of travel is a person when heat radiation or sound of an object is above a low threshold and optionally below a high threshold. In other embodiments, the robot is configured to slow to less than 10 cm/sec when an object in the direction of travel is determined to be a person by heat radiation or sound detection.
In another aspect, the invention relates to a robot for interacting with a resident in an environment, including a sensor adapted to monitor a resident when the robot is in an environment and a resident is within range of the sensor, a controller to monitor surveillance data from the sensor, a transmission control routine configured to enable transmission of surveillance data from the sensor to a location outside an environment, if a resident grants permission for the transmission, and an override routine configured to enable transmission of surveillance data from the sensor to a location outside the environment at least in an emergency condition.
In another aspect, the invention relates to a robot for interacting with a resident in the resident's home, including a sensor capable of monitoring the resident when the robot is in the resident's home and the resident is within range of the sensor, a controller connected to monitor surveillance data from the sensor, a transmission control routine configured to enable a transmission of surveillance data from the sensor outside the resident's home if the resident grants permission for the transmission, and an override routine configured to enable a transmission of surveillance data from the sensor outside the resident's home (i) in an emergency condition detected by the robot or (ii) after a n authorized remote caregiver sends to the robot an authorization previously permitted by the resident and an emergency condition indication, wherein the transmission of surveillance data by the robot outside the resident's home is otherwise prohibited. In certain embodiments of the above aspect, the robot includes a connector capable of receiving a lock-out member, wherein the transmission control routine enables transmission of surveillance data from the sensor outside the resident's home only when the lock-out member is inserted in the connector. In other embodiments, the sensor includes a camera, and the robot includes a privacy controller configured to orient the objective lens of the camera to face the robot's body when a privacy mode is selected. In alternative embodiments, the robot includes a privacy member configured to physically unplug the sensor when a privacy mode is selected.
In yet another aspect, the invention relates to a robot for interacting with a resident in the resident's home, including a sensor capable of monitoring the resident when the robot is in the resident's home and the resident is within range of the sensor, a controller connected to monitor surveillance data from the sensor, a connector capable of receiving a lock-out member, a transmission control routine configured to enable a transmission of surveillance data from the sensor outside the resident's home if the resident grants permission for the transmission and only when the lock-out member is inserted in the connector, and an override routine configured to enable a transmission of surveillance data from the sensor outside the resident's home only when an emergency condition signal is received by the robot.
In another embodiment, the invention relates to a robot capable of human-robot interaction, including a head assembly including a face at least one camera mounted proximate the face, and a positionable support having a first position and a second position, the mount supporting at least one of the face and the at least one camera.
In still another aspect, the invention relates to a robot capable of human-robot interaction, including a head assembly including a head having a face on one side, at least one camera mounted within the face to direct an objective lens of the camera in the same direction as the face, a rotatable pivot supporting the head to permit the head to move such that the face and the at least one camera are hidden from view and the camera is thereby prevented from viewing an area around the robot. In embodiments of the above aspect, the face further includes a matrix panel electronically controllable to show different configurations of left and right eyes composed of matrix elements, and the at least one camera includes a left camera proximate to the left eye composed of matrix elements and a right camera proximate to the right camera composed of matrix elements.
In another aspect, the invention relates to a robot capable of human-robot interaction, including a head assembly including a head having a face on one side, the face comprising a matrix panel electronically controllable to show different configurations of left and right eyes composed of matrix elements, a left camera proximate to the left eye and a right camera proximate to the right eye, each of left and right cameras being mounted near the face to direct an objective lens of the camera in the same direction as the face, and at least one positionable mount supporting the left camera and right camera to each move to a position disabled from viewing an area around the robot.
In yet another aspect, the invention relates to a robot system, including a base station, and a robot, the base station including a wireless transceiver capable of communicating TCP/IP transmissions over a local wireless protocol, a wired Ethernet connector for communicating TCP/IP transmissions over a local wired Ethernet accessing the Internet, and an access point circuit for transferring TCP/IP transmissions between the local wired Ethernet and local wireless protocol limited to a predetermined IP address locked to the robot, predetermined shell level encryption locked to the robot, and predetermined ports to the Internet open only to the robot, the robot including a wireless transceiver capable of communicating TCP/IP transmissions over a local wireless protocol and a client circuit for transferring TCP/IP transmissions over the local wireless protocol. In embodiments of the above aspect, the wireless access point includes a plurality of antennas and an antenna diversity circuit. In other embodiments, the wireless access point includes a plurality of antennas and an antenna diversity circuit and encodes wireless transmissions using orthogonal frequency division multiplexing.
In accordance with an embodiment of the present invention, a method for robotically locating a person being sought may include (a) controlling a mobile robot to activate an annunciator upon an annunciator activation signal, (b) monitoring a sensor suite for any non-noise (noise filtered) response to the annunciator within a limited time period after the activation, (c) comparing any detected non-noise response to a library of sought responses; (d) when the non-noise response does not correspond to a sought response or when no response is identified, identifying a doorway from a present chamber; (e) navigating the robot through the doorway from the present chamber to a new chamber; (f) generating an annunciator activation signal in response to a detected condition, and (g) repeating (a) through (f) until the non-noise response corresponds to a sought response, then (h) activating a found person success routine.
The identifying the doorway may include analyzing robot odometry and heading during wall following and identifying successive odometry/heading combinations indicating a wall end and/or doorway has been traversed. Also, the identifying the doorway may include analyzing robot signal reflection sensing and identifying a heading toward a wall gap. The identifying the doorway may further include analyzing a dynamically built map and identifying a heading toward a wall gap; analyzing an archived map and robot position and identifying a heading toward a wall gap; analyzing a beacon sensor of a door-designating beacon and identifying a heading toward a door designated by the door beacon; and/or analyzing a ceiling configuration or pattern of indicia on a ceiling of a present chamber and identifying a heading toward a door at a position recognized using the indicia or ceiling configuration.
The annunciator may emit a sound including the name of the person being sought, and the comparing any detected non-noise response to the library of sought responses/non-noise response indicates that the source of the non-noise response is a person being sought. The method may further include selecting the second location based on an arbitrary heading and an arbitrary distance for the mobile robot to travel; adjusting the heading of the mobile robot when an obstacle or hazard is encountered; and proceeding until the mobile robot has traveled the arbitrary distance. The new chamber may lie in a different room from the present chamber, and the method may further include navigating the mobile robot from the present chamber location to the new chamber based on one or more of optical room feature recognition, sonar, RFID room tagging, IR directional beacon detection, odometry, inertial guidance, dead reckoning, room mapping, and/or compass navigation.
The method may also include using at least one secondary detection system configured to detect the person being sought, in which the at least one secondary detection system is selected from infrared detection, radiative heat detection, acoustic detection, RFID tag detection, tank circuit detection, motion detection, and/or image analysis; or, the method may further include generating a map of a structure in which the mobile robot is located, and selecting the new chamber based on the generated map, in which the new chamber is in a different room from the present chamber or lies at least a threshold distance away from the present chamber.
The generating the map may include controlling the mobile robot to traverse the robot's environment and to detect walls and obstacles; and the map may be transmitted to the mobile robot by a user. The method may further include locating a closed door, and transmitting an audible signal through the door; in addition, the transmitting the audible signal may include imparting a knocking sound to the closed door.
In accordance with one embodiment, a mobile robot for interacting with a person may include a speaker for generating audible content based on a robot behavior; a scheduler configured to trigger a robot behavior in accordance with the schedule; and a morning robot behavior for interacting with the person soon after the person awakens regarding a second robot behavior scheduled later in the day. The robot may also include a pill dispenser for dispensing medication, in which the second robot behavior includes dispensing medication to the person, and in which the morning robot behavior includes reminding the person about the medication and/or requesting that a person place the medication into the pill dispenser. The robot may further include a transmitter for notifying a caregiver when the medication is not dispensed according to schedule.
Also, the robot may further include a sensor for detecting the person; a locater which uses the sensor when locating the person, in which the robot attempts to locate a person prior to the scheduled robot behavior. Further, each robot behavior may include using the speaker to remind the person of a scheduled robot behavior. The second robot behavior may include guiding the person 2 a predetermined location, in which the morning robot behavior includes verbally reminding the person of an activity associated with the upcoming second robot behavior; and the morning robot behavior may cognitively reinforce the person's recollection regarding an activity associated with the later second robot behavior.
The mobile robot may further include a visual display, in which at least one of the morning or second robot behaviors include both displaying and speaking. The mobile robot may further include a privacy controller for preventing transmission to a caregiver unless the person has given permission to notify the caregiver. Also, the robot may include a receiver for receiving the command from the caretaker, in which the privacy controller is overridden when the command is received.
At least one of the robot behaviors may be input by the person or by the caregiver; and the mobile robot may cease interaction with the person when instructed to stop by the person or by the caregiver.
In accordance with another embodiment, the mobile robot for interacting with a person may include at least one sensor; a privacy controller for preventing operation of the sensor unless the person grants permission for the sensor to operate; and an override unit configured for enabling operation of the sensor regardless of the privacy controller when the override unit receives an override command. Regarding the mobile robot, the sensor may include at least one of a camera, a microphone, a proximity detector, a heat detector, or a tag detector. The mobile robot may further include a wireless transmitter for transmitting and output of the sensor, or a recording unit configured to record the output of the sensor. The mobile robot may also include a teleconferencing unit for transmitting and receiving audio and/or video data, in which the teleconferencing unit includes the sensor, and in which the privacy controller prevents initiation of a teleconferencing session unless the person grants permission. In addition the mobile robot may further include a speech recognition unit for detecting spoken input, in which the permission to talk with the sensor and the input by the speech recognition unit. The mobile robot 1 may control or manipulate an object in the household of the person (for example, turning on a television at a certain time by using an on-board IR transmitter), or point to an object or location at a certain time of day, as components of robot operation.
Also, the override unit may include a receiver for wirelessly receiving the override command from a remote caregiver. The mobile robot may attempt to locate the person when an incoming teleconferencing request is received. Also, the privacy controller may be overridden when the robot detects an emergency condition-for example, if a person appears to be passed out, in danger, sick, or if there's smoke or if a fire alarm has been triggered, etc. The sensor may be conspicuously positioned when operating, and the sensor may become shrouded, covered, or moved so as not to be seen by the person when the sensor is not operating. Furthermore, the sensor may include a camera, in which the camera is placed in an inoperable position when permission to operate is not granted, and in which a lens of the camera faces an opaque surface in the inoperable position.
The mobile robot may further include a behavior privacy matrix stored in a computer memory and correlating a plurality of robot behaviors to a plurality of permission to scores, in which the privacy controller permits a robot behavior when the robot behavior has been permitted, or prevents the robot behavior when the robot behavior has not been permitted.
In accordance with yet another embodiment, a robot for interacting with a person may include an annunciator for audibly indicating the presence of the robot; a beacon for optically indicating the presence of the robot; a sensor for detecting a safety condition; and a controller for immobilizing the robot when the safety condition is detected. The annunciator may continuously or periodically emit a characteristic sound when the robot is moving. The beacon may shine a strobe light when the robot is moving, as well. Also, the sensor may include a camera, an optical sensor, an IR sensor, a key detector, a motion detector, a contact switch, a pressure detector, a proximity sensor, a sonar, a laser based obstacle detector, a noise detector, and/or a microphone, inter alia.
The mobile robot may further include a platform mounted to main chassis of the robot by a stalk, in which the platform and the stalk convey an impression of fragility, and in which the platform extends to the person's waist-level above the main chassis of the robot. The stalk and robot may have an overall shape and/or mass distribution similar to an inverted pendulum. The mobile robot may further include a tipper ring positioned along the stalk which prevents the person from inadvertently colliding with the platform.
The mobile robot may also include a communication unit for sending information regarding the person 2 a remote terminal operable by a caregiver. The communication unit may not send the information if the person has not given permission, and in addition, it may periodically transmit the information according to a predetermined schedule. Also, the communication unit may send the information when the information has changed. Alternatively, for example, the communication unit may retrieve instructions via a computer network.
In another aspect, the invention relates to a method of human-robot interaction between a resident and a mobile robot, the method including the steps of providing a robot comprising a facial display and a sound generating device, executing a viseme sequence of a communication script segment via the display, executing a phoneme sequence of the communication script segment via the sound generating device, wherein the viseme sequence and the phoneme sequence are executed substantially synchronously, and managing the interaction based at least upon an action of the resident and a flexible response regimen of the robot. In one embodiment of the above aspect, the managing step includes an affect behavior for depicting a response of the robot to an event. The affect behavior may include the steps of providing the display on a movable head on the robot, and moving the head from a first position to a second position upon an occurrence of an event. In the above method, the first position may permit viewing by the resident, and the second position is associated with at least one of a surprise response, a joy response, a confusion response, a concern response, and a dejection response. The affect behavior further may include the step of displaying a caricature. The caricature is selected from a set of caricatures associated with a surprise response, a joy response, a confusion response, a concern response, and a dejection response.
In another embodiment of the above aspect, the managing step includes an interrupt behavior for responding to an asynchronous response by the resident. The interrupt behavior may include the steps of receiving the asynchronous response, interrupting the execution of the viseme sequence and the execution of the phoneme sequence upon receipt of the asynchronous response, executing an interrupt sequence, and recovering the execution of the viseme sequence and the execution of the phoneme sequence subsequent to the interrupt sequence. The interrupt behavior may include delivering an error message. Delivering the error message may include the steps of executing an error viseme sequence via the display, and executing an error phoneme sequence via the sound generating device, wherein the execution of the error viseme sequence and the execution of the error phoneme sequence are substantially synchronized. The recovering step may include the step of repeating the steps of executing the viseme sequence and executing the phoneme sequence. The asynchronous may not correspond to at least one of a predetermined set of responses. The above method may also include the steps of comparing a response to the predetermined set of responses, and executing a second viseme sequence and a second phoneme sequence, if the response corresponds to at least one response of the predetermined set of responses.
In another embodiment of the above aspect, the managing step may include a regimen compliance behavior for providing instruction to the resident. The regimen compliance behavior includes the steps of storing a regimen and a compliance schedule associated with the regimen, providing a reminder to the resident at a predetermined event during the compliance schedule, comparing a response from the resident to a set of predetermined responses, and performing an action based on the response from the resident. In another embodiment, the performing step includes providing a medicine to the resident. In another embodiment, the performing step includes providing a second reminder to the resident. The performing step may include communicating with a device remote from the robot.
In another embodiment of the above aspect, the managing step may include a remote communication sequence for communicating between the resident and a third party located remote from the resident. In another embodiment, the remote communication sequence includes the steps of providing a camera on the robot, and changing a status of the camera from a first setting to a second setting upon an occurrence of an event. The first setting may permits viewing the resident, and the second setting may include a privacy orientation of the camera. The method may also include the step of providing a privacy signal to the resident. The event may include a permission denial from the resident. The first setting includes an unpowered condition of the camera, and the second setting includes a powered condition of the camera, and wherein the camera is oriented for viewing the resident.
In still another embodiment of the above-aspect, the managing step includes a navigation manager for moving the mobile robot in an environment. The navigation manager includes a human-seeking sequence, the sequence including the steps of activating an annunciator in a first of a plurality of locations, monitoring a sensor suite for a non-noise response, comparing the non-noise response to a set of predetermined responses, activating a success routine if the non-noise response corresponds to at least one response of the set of predetermined responses, navigating the robot to a next location, and activating the annunciator. The navigating step includes the step of identifying a doorway, which includes at least one of the group of (a) analyzing robot odometry and heading during a wall following mode and identifying odometry and heading combinations indicating traversal of at least one of a wall end and a doorway, (b) analyzing signal reflection and identifying a heading toward a wall gap, (c) analyzing a dynamically built map and identifying a heading toward a wall gap, (d) analyzing a map and robot orientation relative to the map and identifying a heading toward a wall gap, (e) analyzing a beacon and identifying a heading toward the beacon, and (f) analyzing a pattern of indicia on a ceiling and identifying a heading toward a door at a position recognized by utilizing the pattern. In the method, the annunciator emits a sound including a name associated with the resident being sought, and wherein the set of predetermined responses identifies a source of the non-noise response. The method may also include the steps of selecting the next location based on an arbitrary heading and an arbitrary distance for the mobile robot to travel, moving in the direction of the arbitrary heading, adjusting a direction of travel of the mobile robot to avoid an obstacle, returning to the arbitrary heading when the obstacle is avoided, and continuing movement in the direction of the arbitrary heading until the mobile robot has traveled the arbitrary distance. Additionally, the method may further include the step of navigating the mobile robot to the next location based on at least one of optical room feature recognition, sonar, RFID room tagging, IR directional beacon detection, odometry, inertial guidance, dead reckoning, room mapping, and compass navigation. The method may also include the step of using at least one detector configured to detect a person, wherein the at least one detector is selected from the group consisting of an infrared detector, a radiation detector, and acoustic detector, an RFID detector, a tank circuit detector, a motion detector, and an image detector. In addition, the method further includes the steps of generating a map of a structure in which the mobile robot is located, and selecting the next location based on the generated map, wherein the next location is located at least a threshold distance away from a present location of the mobile robot. The generating step includes detecting obstacles or receiving the map electronically from a resident. The method may also include the steps of locating a closed door, and transmitting an audible signal through the door. The transmitting step includes imparting a knocking sound to the door.
In yet another embodiment of the above aspect, the managing step includes a safety manager for controlling contact with a resident. The safety manager initiates a safety sequence including the steps of configuring an annunciator to indicate a presence of the robot, configuring a beacon to indicate the presence of the robot, configuring a sensor to detect a safety condition, and configuring a controller to immobilize the robot when the safety condition is detected. The annunciator emits a predetermined sound when the robot is moving. The beacon comprises a strobe light. The sensor includes at least one of a camera, an optical sensor, an IR sensor, a key detector, a motion detector, a contact switch, a pressure detector, a proximity sensor, a sonar detector, a laser-based obstacle detector, a noise detector, and a microphone. The above method further includes the step of configuring a physical shape of the robot to limit contact between the resident and a base of the robot. The above method further includes the step of configuring a communication unit for sending information regarding the resident to a remote terminal. Sending information includes receiving permission from the resident, which is sent according to a predetermined schedule, upon an occurrence of a change in information, or is exchanged with a computer.
The foregoing and other objects, features, and advantages of the present invention, as well as the invention itself, will be more fully understood from the following description of various embodiments, when read together with the accompanying drawings, in which:
One embodiment of an autonomous mobile robot 10 is shown in
For the purposes of this description, the robot has several categories of functional systems. Each functional system may include components such as sensors; effectors (actuators); communications; and computational control. The robot described herein has several application systems, where an application system is a system for achieving an overall function to be performed by the robot. For a companion robot, there is necessary overlap with the human/robot interaction system, as enhanced human/robot interaction is a primary function. The robot includes systems (privacy protection, interaction systems) to increase user trust and emotional attachment, which has generic benefits by itself and increases the success rate of any and all other applications. The robot maintains a schedule or schedules of health, hygiene, and or need-specific regimens, reminders, messages, and encouragement techniques for associated users.
The robot is responsive to voice command, touch on screen or surface; detection of movement and human presence and (other stuff); and the robot may communicate via voice synthesis, gestures, earcons, text, images, movement, and “ambient” color response. The form of the head and torso of the robot may be designed to provide shrugging, nodding, head shaking, looking away (change of subject) and other gestural cues; the mobility system permits approach and recede motions (personal space and conversational attention cues); and a facial matrix panel may permit a full range of facial expressions. A chest panel may be an interface to a content player/engine. Low-level behaviors of the robot may accommodate interruption by a person. One embodiment may employ a scripting engine that can interrupt or be triggered by behaviors, and an alternative embodiment contemplates that some behaviors, such as self-play, rescue requests, gestural cues, can be executed as concurrent/arbitrated behaviors. The robot may include a wheeled mobility system such as two drive wheels and a passive caster, responsive to short-range object avoidance and hazard management behaviors. The wheeled mobility system may also be also responsive to mid-range obstacle avoidance planning (path planning according to a SLAM generated map) and long-range goal oriented planning.
The head 16 may be mounted on a riser 13, stalk or neck. A turning (or pan) axis AP runs generally parallel to and within the neck. A nodding axis AN is transverse to the pan axis AP and may substantially bisect the head 16. A display axis may be generally transverse to and intersect either or both of the nod and pan axes, and may project normal from the display 26. The head 16 also contains a display 26 used for facial animation, and which may be adapted for videoconferencing, multimedia applications, menu input, etc. As shown in
The head 16 may be generally shaped as a sphere or other solid of reasonably constant diameter (which permits the head to be seated or accommodated in a neck or shoulder cavity with small enough spacing to avoid catching fingers between the head and neck joints or torso).
The head 16 may be articulated via a pan joint 315 including a pan motor and position encoder 314 and a tilt joint or hollow pivot 302 including a tilt motor and position encoder 340. Either or both of the pan and tilt joints may be yoke-type joints, which has particular advantages as discussed herein. In the embodiment shown in
The torso or riser 13 (riser in general, to lift sensors and/or head elements to a height where the perspective of cameras and sensors is useful for obstacle detection and remote teleoperation) may support the head 16 and pan joint 315. The robot 100 includes the majority of the components in the head 16 or mobility platform 12 rather than in the torso 13 as an aesthetic and functional choice. Locating components away from the torso 13 permits the robot to avoid a ponderous trash-can appearance. Functionally, there are several benefits. Heavier components are placed in the mobility platform 12, lowering the center of gravity. The torso 13 can more easily permit variable elevation or extension (pivoted/jointed/folding or telescoping torso parts) if few components are placed there. The mobility platform 12 supports the torso 13, and may be generally triangularly shaped, or may be shaped as a curve of constant width (i.e., Reuleaux polygon with an odd number of sides). Optionally, the robot 100 fits within a generally pyramidal or tetrahedral envelope, with the mobility platform 12 generally extending along a bottom face, the torso 13 generally extending along one or two vertically angled faces, and the head 16 generally positioned at the apex. In this configuration, much of the internal volume of the envelope is open, and may accommodate a carrying tray or a retracted manipulator (not shown). Additionally, the torso 13 can be arranged to extend from one end of the mobility platform 12 toward the horizontal middle so that the weight of the head 16 tends to contribute to locating the center of gravity in the middle of the mobility platform 12. The target height for the center of gravity is approximately 30 cm from the floor (not necessarily at all times if the head can be elevated).
In a particular embodiment, the head 16 of the robot 10 may serve several functions. The head 16 may include the main control board (an example E20 shown in
The rear portion of the head 16 may be smooth and/or translucent, have a rounded or other outer shape, may include indicia or ornamentation, or may incorporate an additional informational display panel. A microphone 32 may be included to collect audio data, receive verbal commands, etc. Preferably, a vectoring microphone array 305 is provided, with multiple microphones capable of locating sound sources and removing ambient noise using time of arrival and other filtering techniques. The display 26 may be affixed to a main printed circuit board (PCB) housed within the head 16. The PCB may include most of the control electronics of the robot 10, or a second PCB may be mounted parallel to the first PCB. The main PCB may include Low Voltage Differential Signaling (LVDS) interface elements (E10 in
The body 12 of the robot 10 may be supported on three wheels 14 that provide stable three-point contact with or without use of a suspension system. The three wheels 14 may be arranged with two differential forward driving wheels and a rear caster, or vice versa; a greater number of wheels could distribute driving and idle wheels differently. A harmonic drive, direct drive, or conventional gear train may be used between drive motors (
The wheels may be alternatively holonomic or omnidirectional wheels. Such systems are disclosed in U.S. Pat. Nos. 3,876,255; 4,223,753; and 5,374,879, the disclosures of which are hereby incorporated by reference herein in their entireties. The robot may alternatively use as few as two wheels with inverted pendulum dynamic balancing (depicted, for example, in
Other components may be included on the robot 10 that provide additional features and functionality to facilitate bonding between the robot 10 and one or more residents or users. “Resident” is used to describe a person that lives in a home with the robot; unless used otherwise herein, the terms “resident” and “user” may be used interchangeably. A resident could be an elderly person or child who requires additional attention or supervision, various members of a family, the family pet, guests to the household identified by residents, etc. To increase the likelihood of bonding between the robot 10 and resident, the robot 10 may include tactile sensors 34 or “digital skin” over any or all of its body 12 or head 16. These tactile sensors 34 allow the robot 10 to recognize when a resident is touching it (and respond appropriately to such touching). The tactile sensors 34 may also allow the robot to recognize when a resident is using a robot not designed to physically assist the resident for support, and appropriately react (immobilizing and sending an audio warning); such a striking function may serve as an emergency shut down, should the robot 10 be malfunctioning.
The head 16 may be mounted to a moveable neck, or a neck may be rigid and unmovable, and serve as a base for the head to move thereon. A combination of both a moveable head and a movable neck is also contemplated. The moveable head 16 allows the robot 10 to interact more naturally with a resident. In one embodiment shown in
In one embodiment, display 26 is a facial segment or matrix panel and is a monochrome, backlit or non-backlit liquid crystal display (LCD) including between about 100 and about 1000 segments individually shaped as portions of facial features on a neutral gray field. The LCD may be a square, rectangular, octagonal, etc., approximately 10-15 cm across, with a permanent pattern of segments. Neither the refresh rate nor the update rate need be rapid. Any subset of the segments can be on (i.e., dark) at any instant, and segments can change or switch state in about 50 milliseconds to about 1 second. The LCD glass is considered part of the head 16, and the display 26 should move as little as possible relative to the head 16.
The display 26 implements an eyes-brow-nose-mouth object display which is animated in synchronization with speech and interaction to express the changing affect and non-verbal behavior of the robot 10. The display 26 may be only backlit during interaction with a resident, if the backlight consumes a more power than desired. Alternatively, the display 26 may be a reflective panel, or backlit by ambient room light filtered in through the (translucent) head 16, backlit by one or more white light emitting diodes (LEDs) and an appropriate diffuser, or any combination of these. The display 26 also may be an organic light-emitting diode (OLED) panel, which may require no backlight (although one may be used) and could consume less power if light-on-dark facial patterns are used.
For a robot 100 provided with the facilities discussed herein, power consumption from two backlights can be half or more of the power consumed by the robot 100 at any time, and can restrict the operation time of the robot to 2 hours or less (conversely, operating without any backlight could increase this time to 4-5 hours).
The display 26 may be constructed using reflective bistable (non-volatile) technology such as electrophoretic, electronic ball twisting, electronic ink or electronic paper displays, and/or hybrid electronic ink/active matrix displays. Electrophoretic or electronic ink displays are made by a high-contrast, low-power technology that is bistable and retains an image when unpowered (e.g., E-Ink™ high resolution matrix Imaging Film or segment Display Cells from E-Ink Corporation of Cambridge, Mass.; or electronic ink/gyricon sheets from Xerox/3M). This type of technology may have a switching time of about 0.5-1 s, allowing for satisfactory display of caricatured facial expressions. Hybrid electronic ink/active matrix displays combine a thin film transistor (TFT) active-matrix back plane, containing polymer electronics-based pixel driving or conventional TFT technology on steel foil, with a front plane of reflective electronic ink as above (e.g., from SiPix Imaging of Fremont, Calif., or Polymer Vision of Eindhoven, NL). Display types operating on different principles, but that display at least some of the identified characteristics (reflective or transparent, very low power, bistable, etc.) are Fujitsu e-paper displays (cholesteric selective reflection film substrate-based color display from Fujitsu, Japan), or NanoChromics™ displays (nanostructured film electrode/electrochromic viologen molecule display from NTERA of Ireland).
When describing the display 26, the term “segment” panel means, in one instance, a panel made up of discretely shaped elements, rather than an X-Y matrix of pixels. The use of the term “matrix” panel means, in one instance, a panel made up of an X-Y matrix of pixels. The potential benefits of a matrix panel are that it can be used as an alternative (or the sole) informational display, and that the faces or facial expressions to be displayed thereon can be completely changed in software, downloaded, or user-designed. A segment panel, however, can be simpler and more efficient; can give extremely smooth outlines for activated segments; and can be more easily adapted to user or community programming of expression sequences.
While the description herein contemplates an octagonal or other multi-sided polygon panel shape that permits covering most of the surface area of a circular panel window 101 of the head 16, an ordinary 4:3 or other rectangular LCD panel may be used in vertical or horizontal orientation. More than one panel can be used within the circular panel window 101 (e.g., three panels horizontally or vertically oriented can fill more of a circle than a rectangular panel). As shown in
Although the robot 10 of
An exemplary single-screen mobile robots are is depicted in
The stalk 124 includes a track 134, along which the head 116 moves 140 (as depicted in
Track 134 may have a top “dead center” position, such that at the topmost point along the track 134, the head 116 may lie flat (i.e., normal to gravity, G, optionally detected by an accelerometer, gyroscope, digital level, etc., in the head 116). Alternatively or additionally, the head 116 may turn over to face a direction opposite to the directions shown in
The robot R2 may include an display 126 on the head 116, which may function as either a touch screen and/or visual output screen, as described above. A speaker 120 may be located on the head 116 for audio output; a microphone 132 or array 132 may be utilized for audio input. Moreover, the robot R2 may include any appropriate sensors 122 or other accessories, such as a finger-cap heart-rate monitor, to provide medical feedback to a remote caregiver or doctor, described below. The head 116 and/or display 126 may include one or more temperature, mass, weight, or other sensors, which may be used to detect properties of items placed on the tray, or to facilitate appropriate use of the tray as a means to move items from place to place (e.g., to detect too hot items, too heavy items, or to detect items placed too far off-center). The head 116 may be suitable for carrying or holding items (such as, for example, food or beverages), and may include reactive, active, clamping, or passive holder(s) or receptacle(s) for food, beverages, and/or medications.
The robot R3 balances on two coaxial wheels (or a ball) using inverted pendulum balancing (i.e., active or dynamic balancing via servo control of the wheels using gyroscopic sensing), and is self righting. Near the top of the robot is a display or teleconferencing screen 18 as described herein. On top of the robot, pointing at the ceiling, a combination camera 28, 29 handles teleconferencing functions and navigation functions. The combination camera may be oriented at the ceiling for following ceiling and/or light fixture and/or door and corner edges, features, indicia, or fiducials, and may be tilted by, e.g., 90 degrees to face forward for teleconferencing or additional modes of navigation. More than half of the weight of the robot is arranged near the wheels, such that the center of gravity of the entire robot is as close as possible to the axis of the wheels or ball. Because the tipper ring 12, at knee level or above, is significantly larger than any part of the bottom of the robot, it may be less likely that the wheels or toe-level portions of the robot will be tripped on. Similarly, the tipper ring is circular and as wide as a person (e.g., 16-24 inches, preferable about 18), and at knee height will encounter most table and chair legs.
The robot R3 may include an inertial sensor near the middle of the height of the robot (3-4 ft.) such as an accelerometer, which may detect collisions. At the top of the robot, a self-righting platform is provided. The platform is also preferably dynamically balanced to maintain a horizontal orientation, such that items carried thereon will not tip or spill. Additionally, object detection and avoidance sensors are positioned at the top of the robot R3. The platform supports a medication dispenser 17. As shown, the medication dispenser is a circular carousel of slots, each slot holding a small paper cup suitable for containing pills or a “Dixie” cup suitable for a swallow or so of water. The dispenser or pill station may be fixed, rotatable manually, or by motor. The dispenser may include a movable D-shaped “purse” handle (not shown) and may mount and dismount from the robot, and or lock to the robot. The dispenser may also fit into a loader, which may be positioned at another location in the home (e.g., in the bathroom, networked, and containing a magazine of pills for loading the dispenser). Loading techniques discussed herein may use a pre-filled (off-site, at a pharmacy, etc.) magazine type loader to load the dispenser. As shown, the dispenser MD or 17 may also be positioned at another location in the home.
No one or more components are necessary. A variant R5 is essentially alike, but includes an extending or telescoping riser 13A (using telescoping barrels, jacks, screws, scissors, or an articulated and/or jointed arm), such that the medication dispenser 17 or other components can be moved about the house with a lower center of gravity. A variant R4 omits the display 26, 18 and dispenser 17, but still includes a navigation camera 28 and many remaining facilities. Such a robot R4, in omitting a display screen, may guide the resident to a set-top box STB and television for vide chat purposes. That is, for those robots which do not have a teleconferencing camera 28, or even for those that do, a set-top box STB may form a part of the robot system. The set-top box STB may be integrated with the robot base station or a separate component privileged on the robot base station network. As noted, the base station (and possibly the STB) are located in the home at the broadband ingress (typically cable, xDSL modem or fiber optic terminal, but also WAN wireless networks). As shown, the set-top box includes a teleconferencing camera directed toward a position in front of a TV upon which the set-top box STB is most likely positioned. Because a normal household television screen is much larger than any screen that can be carried by the robots R3-R7, in a teleconferencing mode (that may be treated as a regiment event for the purpose of scheduling and bringing the resident to a regiment compliance object—the STB and TV—as discussed herein), the robot may bring the resident to the STB and TV for a videoconferencing session that takes advantage of a large screen (on the TV), stable acoustics (a microphone array on the STB that is not subject to motor and mechanical noise), stable AC power, a stable camera, and wired broadband access.
Two further variants are shown in
The mobility system of the robot R7 of
In accordance with one example configuration, the robot 100 includes six infrared rangefinders for obstacle and stairwell avoidance, with three bump switches for obstacle avoidance when backing up.
The infrared sensors preferably have a similar electronic implementation such as, for example, a synchronous reflectivity emitter/detector pair comparable to cliff detectors. With regard to the infrared rangefinders, as non-limiting implementation examples, emitter/detector pairs may be deployed in accordance with several different configurations as follows: Convergent-Crossed emitter/detector pairs 965 are emitter/detector pairs providing a specific detection zone at the point of intersection, as used in cliff sensors. Diffuse-Parallel 966, 962A, 962B are colocated emitter/detector pairs preferably providing a 30° field of view (+/−15°), detecting objects at distances between (for example) about 2 cm and 20 cm, depending upon the reflectivity of such objects. The Tangential sensor 961 may utilize a “broken beam” topology created across a concave surface 978 at the front of the robot 100.
In accordance with at least one configuration, as non-limiting examples, the sensors can perform synchronous detection during and after illumination pulse of associated emitter; in which each sensor reading produces two scalar values denoting ambient reflectivity and illuminated reflectivity. With a least 8-bit dynamic range, each sensor can produce at least 100 such samples per second. In a preferred configuration, analog conditioning/filtering/integration is done on each detector's signal before moving into the digital domain. Further, the detector components may be standalone photodiodes, or may integrate transimpedance amplifiers to improve the signal-to-noise ratio through a wiring harness. A robot 100 including 16 such sensors is shown in
A preferred configuration relates to a robot including six forward-facing diffuse sensors 966, spaced 7 cm apart, which can provide sensor coverage for objects up to about 14 cm away, provided that the objects are reflective enough to be detected at that range. Three downward-facing cliff sensors 965 are placed before each wheel 14 in its direction of travel. Two angled diffuse sensors 962A, 962B and 963A, 963B at the rear-angled sides of the robot mobility base 12 provide obstacle detection in the reverse direction, and can guide the docking process in conjunction with matching emitters on the dock, for example. Two angled convergent sensors 966 at the front provide a defined detection zone just outside the projected envelope of the robot, for use in wall following and wall locating, for example. A tangential beam across the front 978 of the robot base provides precise contact information when an object touches any part of that surface, and may accordingly accelerate collision detection and accurately locate the object's position.
Because the mobile robot 100 functions in a household environment, the issue of household safety may be addressed by any of a variety of features. Among these features, inter alia, are personal contact avoidance, obstacle avoidance (including the ability to detect and/or avoid objects that exist over a wide range of possible heights and positions relative to the robot 100), and damage avoidance—avoiding and/or mitigating damage to either the home, objects in the home (such as furniture, fixtures, and the like), or the robot itself, for example. In connection with these features, the robot preferably includes redundancy in detecting and/or avoiding obstacles.
Regarding personal contact avoidance, the robot 100 preferably avoids initiating physical contact with any person—such as the user—unless there are circumstances that make such contact unavoidable or necessary. However, as non-limiting examples, the robot 100 may initiate contact when a possible medical emergency requires the robot to measure the temperature of the user, or when an authorized remote user issues a command to the robot to make physical contact with a person, inter alia.
In accordance with at least one aspect, the robot 100 includes one or more sensors for detecting the presence or absence of objects over a range of physical space, in which the monitored space is divided into protection zones. As illustrated in
As illustrated in
In the area in front of the robot 100, the light curtain may detect incursions by objects at a height between approximately one through 12 inches above the ground. Toward the face 26, the near zone Z30 may also be monitored by an infrared or ultrasonic obstacle sensor. At far range, such as two through twenty feet or more, the robot 100 may include a camera 304, which can detect both near and far objects when image processing is applied to the data produced by the camera 304.
In order to detect people in front of the robot 100, heat sensors, ultrasonic object sensors, or voice vector detector arrays may be used. For example, the robot 100 may include an array of six microphones (not shown) positioned such that the direction of the source of a sound may be determined; when the robot “hears” sound that corresponds to speech and the voice vector array indicates that the source of the sound is in front of the robot, the robot may respond accordingly. In addition, the robot 100 may employ such “people detectors” at either side and/or behind the robot 100, in order to detect people who may be standing or approaching too close to the robot (such that a danger of collision or of the robot tipping), in order to respond appropriately.
The robot can receive data from two or more sensors, integrate the sensor data to predict the likelihood of an obstacle being present in one of the protection zones, and modify its actions (for example, by reducing speed, changing the direction of travel, and/or lowering the head 16, inter alia) when the predicted likelihood of an obstacle being present exceeds a threshold. Moreover, the threshold can be adaptively modified in accordance with experience—for example, when the robot collides with an obstacle after failing to predict the presence of that obstacle, the robot can lower the threshold; or when the robot discovers that a predicted obstacle did not in fact exist, the robot can raise the threshold.
As an advantage, by predicting the presence of an obstacle coming into the path of the robot based on input from more than one sensor and/or more than one protection zone, the robot can avoid or mitigate obstacles encroach onto a primary protection zone.
As illustrated in
In addition, the robot 100 may include a microphone that monitors ambient sound or noise levels. When the microphone input indicates sound corresponding to speech, for example, the robot may determine that it is likely a person is in the vicinity, and accordingly reduce its movement speed, in order to reduce the chance of colliding with the person. The robot 100 may also monitor other sensors such as heat detectors (infrared photosensors, for example) or cameras, and make behavior determinations based on one or more of the sensors. For example, if no person is in fact in the vicinity of the robot, but the television is turned on, then the robot's microphone detects speech, but the heat detector does not indicate the presence of any heat sources. In such a case, the robot 100 may instead determine that no person is present, and accordingly decide not to reduce its movement speed.
When the robot 100 is in an environment containing many potential obstacles (for example, in a cluttered basement or in a room full of people), the robot may operate at slow speed in order not to collide with people or things. When operating at slow speed, the robot may rely more on bump sensor input than on optical cameras or photosensors, for example.
The cameras on the robot, depicted for example, as item 28 in
If utilized, a shield may be transparent to allow clear viewing of the camera 28 within the hollow 24 a. Alternatively, it may be entirely opaque and movable into a position to cover the camera 28, so the resident may directly control the ability of the robot to view and potentially transmit visual data. In one embodiment, the camera and/or lens is entirely withdrawn into the robot's body, or the lens is turned to abut the robot's body.
Accordingly, the robot is not only rendered incapable of recording its surroundings, but can be readily confirmed as incapable of such recording, as indicated by the blocking of the lens, the turning of the lens to substantially face the robot's body, or the turning the head to change the field of view. The visual cue of camera or head position may be supplemented by an electronic light that may be dimmed or undimmed when the camera is recording visual data. Thus, the person can clearly see and confirm that they are not under potential surveillance by a third person. A similar mechanism can be used with a microphone 305 or microphone arrays 305 on the robot; the robot may disconnect, shutter or otherwise disable microphones when entering a privacy mode. If the microphone(s) 305 are turned along with the head or otherwise moved to be blocked, an elastomer seal in the inoperative position may further muffle sound. Similarly, external antennae can be disconnected. In each case, the transmitting medium is rendered incapable of monitoring the resident.
As shown in
In an alternative model, the key slot, plug, or other connector 321 does not mechanically close electrical contacts, and can be overridden in emergency situations. In this case, software transmission control routines may enable transmission of surveillance data from the camera 28 outside the resident's home only when the lock-out member is inserted in the connector 321. In this model, the robot 10 may normally use its controller connected to the camera or microphone to monitor surveillance data from that sensor. The robot may run a transmission control routine configured to enable transmission of surveillance data from the sensor (to the caregiver)outside the resident's home if the resident grants permission for the transmission. However, an override routine may be configured to enable a transmission of surveillance data from the sensor outside the resident's home (i) in an emergency condition detected by the robot. Alternatively, if an authorized remote caregiver sends an authorization previously permitted by the resident together with an emergency condition, the camera and/or microphone may be enabled. wherein the transmission of surveillance data by the robot outside the resident's home is otherwise prohibited. This kind of arrangement may also use a privacy member configured to physically unplug the sensor when a privacy mode is selected, if the override mode uses a mechanism or relay to “replug” the sensor.
In accordance with at least one embodiment, any sensor input capable of observing or monitoring the environment surrounding the mobile robot may be routed by a privacy device 350, as illustrated in
An override unit 354 may override privacy directives issued by, or monitoring capabilities disabled by, the privacy controller 352, and thereby permit the sensor input to be temporarily recorded and/or transmitted, despite the permissions implemented by the privacy controller 352. In certain embodiments, the override unit 354 may only override the privacy controller 352 when either a predetermined event is present, the event being either a threshold on the output of one or more sensors or interpreted data as detected by the controller 352, or when a caregiver's override command is received by the mobile robot. An override command from the caregiver may be in the form of an encrypted code or other secure transmission, encoded or digitally signed using any suitable method, such as, for example, digital certificates, public-key encryption, a predetermined password, or any other suitable mechanism.
The override unit 354 may permit surveillance or monitoring temporarily when, e.g., a distress condition of pulse, breathing, medical, environmental, or safety monitoring sensors, or other emergency situation is detected. When a distress condition is detected, the robot may also make an outbound call requesting the caregiver to check on the resident. Also, the mobile robot may include a receiver for wirelessly receiving the override command (from a remote terminal, which may be operated by a caregiver, for example). In addition, there may be circumstances in which considerations of safety or health suggest that the usual permissions granted or denied by the resident be temporarily overridden. Master override permissions may enable a caregiver to receive permission to temporarily override subsidiary permissions in emergency situations, or to enable medical monitoring devices or data analysis conclusions to override these permissions.
Circumstances that may require overriding of the privacy controller 352 may include a determination by the mobile robot that the resident has become non-responsive (and therefore possibly unconscious), or when a fire alarm or smoke detector has triggered on the robot, or in the environment. In such circumstances, the override unit 354 may override the privacy controller 352 to permit input from the camera 28, microphone 32, or other sensors 22, to be transmitted to a remote terminal belonging to a caregiver, police or fire department, or ambulance service. Therefore, the caregiver may be alerted to such emergency circumstances, and receive relevant video, image, or audio information from the mobile robot in proximity to the resident.
Higher level planning is not always necessary, but can be handled by specific controls that influence or direct the behavior based systems. As shown, the behavior bases systems normally intervene between planning routines 6A-2 and actuators 6A-4. Long-range navigation, as discussed herein, will move the robot from room to room as discussed herein, or alternatively using path planning. A mapper may build maps, which may or may not be “Cartesian” spatial maps (most of the routines, techniques, and inventions discussed herein do not rely upon Cartesian or spatial maps). Stereo feature extraction may proceed according to known routines (e.g., SIFT or SURF scale invariant feature extraction) in order to identify landmarks or labels. Some of the necessary records may be stored remotely on a server or database 6A-3.
In the example of
The lower PCB board E46 includes one or more power supplies E90 and a battery and/or charger E91, in which a power supply channel (which may be a plastic-sheathed copper wire, or any other component suitable for power transmission from the mobility base 12 through the torso 13 and into the head 16) transmits power from the power supplies E90 to the middle PCB E30 and to the upper PCB E20. The lower PCB E46 also includes a pan motor amplifier E401 and one or more drive motor amps E402, together with encoder E403 and proximity sensor interfaces E404, and also a two-dimensional accelerometer E406 and a microcomputer E407 (which may include a microcontroller, a programmable logic array, a digital signal processor, and/or a general-purpose CPU, as non-limiting examples). The lower PCB E46 communicates with proximity sensors 222 (routing such communication to the proximity sensor interface E404), drive motors/encoders 46, and to the pan motor/encoder 314.
Regarding the torso area, various features local to this region—such as the body display panel 18 and speakers S31—communicate with the middle PCB E30. Also, an array of microphones may be disposed on the robot 100 for determining the location of the source of a sound. In accordance with at least one embodiment, a microphone for use in a sound-source locating system (herein termed “Acoustic Magic”) preferably transmits a continuous serial stream of octets indicating the direction of current sound. The stream is transmitted by a DSP onboard an Acoustic Magic PCB, through its synchronous serial port, emulating an asynchronous serial signal.
The serial signal is received by an HCS12 on the power board, and the octets may be delivered unprocessed in the packet stream to the CPU board.
On the CPU board, a process receives the signal and decimates it to (as a non-limiting example) 10 samples/second, storing the stream in a shared-memory ring buffer. The decimation is preceded by a single-pole IIR low-pass filter, with a pole at 3 Hz, for example. The buffer preferably records at least the last 5 seconds of observations, each of which are time-stamped with a true timestamp for when that audio sample was actually heard (accounting for any buffering through the signal chain).
The estimated voice location is produced by combining this vector with the speech detector of the recognition system. The speech detector preferably may issue regular callbacks indicating its live status. These callbacks correlate detected speech (compensating to true real time) with the 10 hz voice location samples, to produce the estimated direction vector to the most recent detected utterance. These utterance direction vector samples are similarly recorded in a shared memory ring buffer, for use by the user-location and interaction system. Such functionality and processing is preferably provided by an array processor E301 disposed on the middle PCB E30.
In the head area, various sensors such as a pyro sensor S21 or touch sensor S22 may communicate with the upper PCB E20, as well as the right and left image sensors (cameras, etc.) 28, 29, and the face matrix panel 26. The tilt motor/encoder 340 also is connected to the upper PCB E20.
Within the upper PCB E20, components corresponding to devices disposed in the head 16 may be integrated onto the PCB board, such as: a face controller/interface E210, for interfacing with (and/or controlling) the face matrix panel 26; audio codecs/amplifiers E211; and/or Ethernet E202, USB E203, or infrared (IR) E204 interface components for communicating according to their respective channels. Further, the upper PCB E20 may include an embedded microprocessor E230 (preferably consuming less than 1 Watt per hour) and a memory E209 comprising volatile and non-volatile sections, for providing necessary instructions to coordinate and/or control the various subcomponents on the upper PCB E20.
In addition, the upper PCB E20 in the head 16 may be connected to an antenna E201, such as a cellular-band antenna for use with a CDMA modem, in which the antenna E201 includes an embedded antenna component mounted on a PC board. As an advantage, such components are designed for cellular handsets, are small, precisely manufactured and impedance-matched, and easy to integrate. Further, the antenna design may alternatively use simple wire dipoles to additionally reduce cost.
Preferably, a dipole design is implemented directly on the upper PCB E20 in the head 16. One example cellular modem slot supports modules operating in the 800 mHz cellular band, specifically the transmit band from 824-849 mHz and the receive band from 869-894 mHz. The robot's head 16 may be selected to be large enough to accommodate a half-wave dipole for these frequencies, which is nominally 163 mm long. However, if many other mechanical and electrical components are packed into the head, the antennas may alternatively be mounted horizontally. As another alternative, an “inverted V” configuration may be used, which is a horizontal dipole, extended slightly (by approximately 4%, for example) and bent in the middle. The inverted V has the added advantage of a more uniform omnidirectional response pattern.
A preferred configuration places two “inverted V” dipoles, mounted at the top and bottom of the head for spatial diversity. Each dipole has two legs, each preferably 85 mm in length, 2 mm in diameter. The calculated VSWR for this configuration may range from 1.7 at 824 mHz, 1.2 at 860 mHz, to 1.7 at 894 mHz. These antennas may be implemented with (for example) 12-gauge wire segments affixed to Plexiglas supports placed behind the CPU board, in which case 50-ohm coaxial cables can connect the antennas to the CDMA modules.
Step SD2 represents a first interaction, which, depending on application, may be a greeting to the resident, an audible comment to itself, etc. At Step SD3, the robot may ask the robot a question or make an affirmative statement, “How was your day?”, “Got plans tonight?”, “You asked me to check movie times for you.”, etc. The question be designed specifically to elicit one of a specific number of responses from the resident, or the resident may respond to an open-ended question or statement. A Response X (outside the robot's library of anticipated responses) may cause an error, confused, or other reply from the robot, Step SD, causing the robot to end the interaction or pose the question or statement again, Step SD3. Predetermined or expected responses may initiate corresponding replies, Step SD5, from the robot. These replies may include a matching affect responses as well; e.g., an audible “Wow” with an affect indicating surprise, an audible “That's too bad” with an affect indicating concern, “Sounds great” with a happy affect, etc. Alternatively, the robot's response may be a second question. For example, if the robot asked the resident which movie the resident would like to see, the robot could then ask if the resident if the robot should search for theaters showing that movie, and so on.
After the robot's reply at SD5, the robot may begin a number of second interactions, at Step SD6. For example, the robot may begin playing a game with the resident, move on to a second, more specific line of questioning or interaction, or end the interaction in view of particular responses. Certain interactions may lead to other interactions. Under these circumstances, the robot may interact with the resident for long periods of time, changing topic, activity, or task, as the circumstances demand, creating a rich interactive experience for the resident.
According to these routines, the robot may conduct an effective method of human-robot interaction. Most fundamental motion behaviors are asynchronous (or independent of) the dialogues. Behaviors such as escape or safety reactions to events are not affected by the dialogue. The communication script segments (queries, choices, replies) are received by the robot from its internal database or a remote database, and include the dialogue branches with robot speech prompt text (such as dialogue queries, replies, and confirmation/feedback responses) as well as human response text (such as the choices for the person to repeat or select). The robot interprets the communication script segment to generate an audible robot speech prompt, either by referring to recordings; by consulting a speech synthesis phoneme script that goes with the text; or by having on-the-fly text-to-speech synthesis (each of these being capable of including script or timestamp synchronized viseme tags defining facial animations, earcons defining emphasis sounds, or “animotion” scripts defining robot emphasis motions). Animotion scripts tend to define a desired expression during the communication script segment by adding an expression motion, usually one or both of a head movement sequence including nod axis head movement or turn axis head movement, or a robot movement sequence including movement of the entire robot.
The robot receives input in the form of either recorded speech from the person via its microphone array, or the result of a push-button or touch-screen identification of the selected response choice. If at any point during this dialogue, the robot detects an event that requires a response from the behavioral system (e.g., in moving into an expression motion, the robot bumps a table or other obstacle, or has moved partway toward a second person), the robot interrupts the dialogue branch in response to the event detected by one of the plurality of behaviors (i.e., via the sensors used by those behaviors) to execute an asynchronous response (such as stopping the robot). After any appropriate asynchronous response (including a further comment from the robot regarding the change its motion or pose, or what was discovered that caused the change and asynchronous response), the robot recovers the dialogue branch after the execution of the asynchronous response. The robot interrupts expression motion(s) in response to the event detected by the robot. The robot is stopped (or other corrective behavior) to reposition the robot according to the event.
Subsequently (step 7D-16), the robot attempts to receive an audible response and process it.
As shown in
In this manner, the received (from database) communication script segment includes an output query sub-script text and a response tree of five or less sub-script response text candidates (often “yes”, “no”, and an alternative such as “I don't know”; on other occasions a list, e.g., pill type “red diamond”, “green”, “white”, “none of the above”). The list is kept at five or less so as to (1) avoid overwhelming attention; (2) avoid overwhelming display screen “real estate,” (i.e., no scrolling, text large enough for easy review) and (3) limit the resources necessary for speech recognition. The output query text (i.e., the questions asked) are related to or associated with an audible output signal (e.g., phoneme codes, compressed recorded audio), and output as a spoken query to a person. Both the output query sub-script text and the limited number of responses are displayed together on a display of the robot. The robot receives, via a microphone or a vectoring or other microphone array, the person's spoken response as an audio input signal, which is processed to check/recognize if it corresponds to what has been displayed (i.e., all the responses expected to the question). If the signal recorded by the microphone is recognized, the robot issues an output signal (i.e., a voice or text confirmation, highlight, or blinking) to feed back or repeat back the recognized response. If not, an output signal, also a voice or text reprompt, highlight, or blinking prompts the user to retry communicating a response. In some cases, the robot can request confirmation from the resident by monitoring for a confirmation of the correctness or incorrectness of the recognized choice.
In the event that audio reprompting is ineffective, the reprompt can take the form of highlighting (including blinking or the like) the displayed sub-script text on the display of the robot, and at the same time receiving the response choice via a manually operated control (e.g., soft button, touch screen) associated with the display. There is some advantage if two of the response text candidates are a simple affirmative and simple negative response, and if the processing then checks for speech corresponding to many different kinds—a family—of affirmative responses equivalent to the simple affirmative response (yes, yeah, uh-huh, sure, etc.) or includes speech corresponding to a family of negative responses equivalent to the simple negative response (no, nope, nah, I don't think so, uh-uh, etc.). Other messages can be common among many scripts (I don't know, please go back, etc.), and in such cases, the robot may recognize a limited set of possible responses that are not shown on the screen (please go back, come back later, quit, etc., show me my schedule). If there is no yes or no answer, the list of answers may be limited to three or less to significantly limit speech recognition work.
A standalone network adapter provides firewall circumvention and zero-configuration of network access for the robot, without requiring a user-interface for configuration. This may be done independent of the robot and, in certain embodiments, can potentially bridge multiple robots in the home. In one embodiment, WiFi is an non-broadcast extended service set identifier (ESSID), with wired equivalent privacy/WiFi-protected access (WEP/WPA) enabled by default and “locked”. The robot and network adapter may be paired at the factory with generated, unique security information. Additional pairing operations performed at a later time may require physical proximity or line of sight between the robot and network adapter (as verified by pairing cable, by short range RF such as BlueTooth, ZigBee, Wireless USB, or by short range IR.
The network adapter functions as a firewall and may be into outside any existing infrastructure. It also includes “Simple Traversal of User Datagram Protocol (UDP) Through Network Address Translators (NATs),” as defined, for example, in http://www.faqs.org/rfcs/rfc3489.html, the disclosure of which is hereby incorporated by reference herein in its entirety. This functionality may be embedded into the device, thus allowing the network adapter to act as an Internet gateway, granting Internet-access to recognized robots in a home environment. There are a number of ways to access the firewall circumvention features of network adapter, for example, by the robot having a MAC address within a known range.
General enabling network architecture is shown in
The base station (in this view, 8A-6) is always connected via a modem or network terminal (usually a cable modem, xDSL, or optical network terminal) to a (usually broadband) Internet Service Provider (not shown) and then to the public internet (cloud). The robot base station generally has the functionality of and provides a router (e.g., DHCP server, NAT, firewall, etc.) so does not need to traverse a router. Nonetheless, as described herein, in some cases the robot base station 8A-6 will be behind a firewall or NAT service, and some techniques for matchmaking are addressed herein.
The robot base station 8A-6 regularly or intermittently connects to a matchmaker 8A-10 (a web or database server), which maintains up to date records of the current IP address of the robot base station 89A-6 (assumed to be, in most cases, dynamically assigned). Once a match is made (i.e., a caregiver station 8A-4 and robot base station 8A-6 are informed of each other's IP address), a secure peer-to-peer connection (as shown in
In addition, the matchmaker 8A-10 may be part of a network which relies upon so-called supernodes to provide routing among privileged clients, or the matchmaking function may be handled by a network of nodes and supernodes which do not have any access to the data network 8A-8, 8A-12, 8A-14. For example, supernodes may be nodes with a public or fixed IP address, not behind a firewall. Cryptographic keys for user identity and password are stored on the matchmaker 8A-10 as a login server. Each robot base station or caregiver station may use a protocols similar to STUN (rfc 3489) protocol and TURN protocols to determine what kind of NAT or firewall intervenes. To initiate, the station may open (TCP or UDP) listening ports at a random port number and specific common port numbers (80 for HTTP and 443 for HTTPS). Each station may maintain a cache of supernodes or matchmakers and their IP addresses and known random ports. Upon connecting to the internet, either station may send packets (TCP or UDP) to a supernode or matchmaker IP address at the historic port number, await a response, then try to connect on the HTTP port and then HTTPS port. After being connected to the supernode or matchmaker, this matchmaker or another matchmaker, Web server or security server may authenticate the station. The supernode or matchmaker may also exchange sufficient packets to inform the station of the IP address of a different matchmaker or the address of a particular login authority that can authenticate that station. During a connection, each station may populate its cache with supernodes, nodes, matchmakers and their last known (random) open ports and IP addresses. Any node (robot base station, caregiver station) or supernode may act as a router to forward and route traffic between a matched robot base station or caregiver station if either or both are behind port-restricted network firewalls or NAT. This is typically similar the “Skype” peer to peer telephony network, and would be capable of interoperating with such a network. In the present arrangement, preferably, idle caregiver stations (typically ordinary PCs running authorization software, VPN, and/or client software) act as routers in such cases, as this would tend not to burden the supernodes, and also would avoid using robot base station nodes which would typically have less processing capability (i.e., about the same as a consumer router).
As shown in
The base station 8A-6 is also connected to a Web server or server 8A-8 which hosts various applications useful on the robot 10. The Web server 8A-8 has access to a file system 8A-12 with content to be used by the robot 10 (e.g., multimedia content); and also to a generalize database. The database may store user profiles for the residents, the dialogues available for robots 10 to use, schedules and the like. Any data considered confidential or sensitive can be obscured as to identity or content by cryptographic keys. Data may be used or processed by the database in a manner that the database owner (not the resident) may not inspect or attach an identity to private date. Join tables can match records stored on the robot or base station with data stored by the caregiver, so that data collected or sent from either while the other is not on-line to connect peer-to-peer 8B-8 may be associated. Data normally preserved only on the robot or by the caregiver or caregiver's network (confidential medical history and records, etc. may be maintained on the database 8A-14, but preferably in an encrypted backup form that is recoverable only by one of the caregiver, robot, or robot base station or a proxy (e.g., the resident's home PC).
In this manner, a robot system, includes both the base station 8A-6 and robot 10. The base station includes a wireless transceiver capable of communicating TCP/IP transmissions over a local wireless protocol and a wired Ethernet connector for communicating TCP/IP transmissions over a local wired Ethernet accessing the Internet. The access point circuit (e.g., in the robot base station) transfers TCP/IP transmissions between the local wired Ethernet and local wireless protocol, and is limited to a predetermined IP address assigned to and locked to the robot (e.g., cannot be changed by a resident or caregiver without authorization, or at all), predetermined shell level encryption locked to the robot (such as the discussed VPN and/or SSL), and predetermined ports to the Internet open only to the robot, (e.g., one random port assigned to each robot and opened, or under the same language, a common port such as HTTP or HTTPS in combination with an intervening random routing node). The robot itself would include a wireless transceiver capable of communicating TCP/IP transmissions over the local RF or wireless protocol and a client circuit (e.g., CPU and OS with IP protocol stack) for transferring TCP/IP transmissions over the local wireless protocol.
Preferably, the robot base station includes a plurality of antennas (the robot may also), and makes use of 802.11n multiple-in, multiple-out antenna diversity and selection to overcome issues of multipath propagation within the home, not limited to 2.4 GHz unlicensed frequencies, but in bands from e.g., 900 MHz up to around 10 GHz. Further preferably, the robot base station further encodes wireless transmissions using orthogonal frequency division multiplexing. In such a case, the robot base station may be portable and incorporate elements of the caregiver station (such as a screen capable of relaying camera and video information for teleoperation, and a controller for teleoperating the robot), and the embodiments contemplate that the platform for the robot may be treaded and skid steered, with surveillance cameras on board. Video is transmitted from the robot to the base station at the highest possible speed, multipath resistance, and frame rate using multiple-in, multiple-out antenna diversity, orthogonal frequency division multiplexing, IP or IP-compatible packet protocols that do not error correct (like UDP), and optionally predictive wavelet based video compression like H.264AVC or MPEG 4.
In the embodiment depicted in
A central service may provide a matchmaking similar to dynamic DNS services, but for commercially available robots that utilize the particular network adapters. Under these circumstances, each robot would be shipped to a consumer with a unique pass code to map the owner to the device when the owner first connects to the matchmaker service. In this case, the matchmaker service is web-hosted centralized user interface for robots. Service proxies configuration and user interfaces between robots and users, who are both Internet clients. Additionally, the matchmaker service may operate as a fleet management tool (for software updates) and a portal for the robots in a particular home. From this central site, a user can “connect” to a robot, which redirects the user to the user interface hosted on the robot.
In response to schedules or events throughout the day, the mobile robot generates updates including updated information regarding the resident, including but not limited to: (a) medication compliance information; (b) presence or absence of the resident, recognized by patterns of sound or other detections; (c) status or “nothing unusual” reports; (d) reports of visitors; (e) updates entered by the resident as to their planned schedule (“I'll be out for two hours”); (f) report of unusual patterns of sound (distress noises, calling out noises) or unusually loud sounds; (g) requests for communication from the caregiver or other family members to the resident. These updates include event-responsive updates, data-change updates, status updates, and scheduled updates.
The robot uploads the update information to the computer server, which is integrated with or in communication with a push server, using a push updating system. The push server, similar to a mail delivery agent, sends (or pushes) newly received updates to the update client, similar to a mail user agent. Both the robot and caregiver's terminal may be an update client. The push server monitors the network of robots and caregiver clients, and when it sees a new update for a caregiver, it retrieves the update and pushes it to the caregiver's terminal. Similarly, when the push server sees a new update for a robot, it pushes the message to the robot. The updates may or may not be separately maintained and archived.
As shown in
The robot of the present invention provides a rich interactive experience for a user by utilizing visual cues to indicate its responses to certain events. Those cues include, at least, image displays on one or more of the screens and head or other physical movements to convey reactions.
As displayed, mouth, eye, and eyebrow objects (such as those exemplary expressions depicted in
The following discussion describes either a segment panel or matrix panel, with distinctions noted at appropriate portions. In discussing the matrix panel, terminology used to create animations in Macromedia™ Flash™, a portable animation system, is useful. In this terminology, a “movie” may be a collection of scenes, which is a collection of objects such as text objects and image objects, which in one embodiment may be placed in layers onto a background. The movie and objects are responsive to events, and may apply effects (animations that change the appearance of an object) as well as carry out actions outside the movie. In discussing the operation of a segment panel, an object may be a single segment or group of segments that may be turned on or off in sequence or separately, which can give the impression of motion or a change in appearance. In both cases, “objects” may be animated and accommodate transitional animation states. Notwithstanding the use of this terminology, the invention is not limited to the discussed animation systems.
The display 26 implements an eyes-nose-mouth display which is animated in sync with speech and interaction to express the changing affect and non-verbal behavior of the robot. The animated expression of the display is controlled subject to a number of variables within the system: non-verbal behavior synchronized with speech, phoneme animation synchronized with speech, base expression from dialogue and affect module, and several specialty expressive sequences. The functional elements of the expression include several base expressions, several eye gaze directions, several eyebrow positions for emphasis gestures, and several mouth shapes (visemes) for phoneme animation. These elements are independent (orthogonal), i.e., phoneme animation, gaze shifting and eyebrow raising work regardless of the base expression, which include Sad (mild), Concerned, Neutral (positive), Happy (with at least 4 gradations from contentment to joy), Surprise (with different presentation for positive and negative surprise), etc. There are “in-between” states (“tweens”) in all these elements, for smoothing animated transitions from state to state.
Eye gaze direction may provide, for example, 16 different positions of direct gaze from right to left. Additionally provided may be two positions upward-left and upward-right, and two positions downward-left and downward-right. Reflection-indicators on the eyes may be omitted, set depending on directionality of room illumination, or kept stable. At least three eyebrow positions may be provided —neutral, raised medium, raised high, plus tweens. Eyebrows participate in forming base expressions as well.
Phoneme animation utilizes a subset of 13 visemes (a “viseme” is set of animation frame frames of a mouth in positions corresponding to speech synthesis phonemes, according to the timing and pace of human mouth motions used to form such phonemes; an “earcon” is the audible equivalent of an icon, i.e., a short, meaningful sound that conveys an idea, object, or action without the use of spoken language), tweened for target speech rate of 5 phonemes per second. Expressions that indicate a tongue may be eliminated to reduce this subset. A set of closed, A, B, E, F, K, O, OO, R, TH comprises a basic set of phonemes. A larger set may include any or all of Silence; P, B, M; W, UW; R; F, V; TH, DH; L; S, Z, D, T, N; SH, CH, JH, ZH; Y, IY, IH, IX, AW, H, K, G, NG; EY, EH, UH; AE, AX, AH, AA, AO, ER, AY; OY; and OW.
Specialty expressions may include Wink, Blink, Curious, Inquisitive (e.g., one flat eyebrow with sideways glance), Dilated pupils, Laughing, Crying, Sleepy (used when the robot battery level becomes low), Thinking (used while downloading something, might include brow furrowing). The expressions depicted in
As generally depicted herein, the form of the head and torso of the robot may be designed to provide shrugging, nodding, head shaking, looking away (change of subject) and other gestural cues; the mobility system permits approach and recede motions (personal space and conversational attention cues). It is important to note that it is not necessary that the robot have a head, or indeed any human or animal attributes. For those known gestural cues that involve motion of a person's head, shoulders, or the like, an analogous motion using portions of the robot that are in similar positions, and/or of similar profile, and/or are exaggerated using the entire body of the robot, if arranged according to the same or similar conversational timing and cues as human gestural cues. Some of these are shown in
In accordance with an embodiment of the present invention, a mobile robot may provide interaction with a resident to promote or insure compliance with various tasks or activities. For example, the mobile robot may provide reminders and other stimuli for numerous scheduled events, as illustrated in
As shown in the schedule 50 in
Additional regimens may be supported by the mobile robot, with similar reminder schemes. For example, the robot may provide the resident with reminders for social visits 56, or entertainment events 58 previously placed in its schedule 50. During the medication compliance routine 54, the robot may remind the resident of the purpose of the interaction, note the amount of medicine that must be taken, and request permission to proceed with the compliance routine. In the present example, the mobile robot may, depending on whether the robot includes a medicine dispenser, magazine, or carrier, (i) direct the resident to go to the room in which medication is kept, (ii) offer medication from an cup or cups borne by the robot, along with water, and/or (iii) open an appropriate portion of a pill dispenser, in order for the resident to retrieve and take the medication. Accordingly, by simply reminding the resident that the time for regimen compliance has come, the probability of regimen compliance is increased. However, various actions can be taken by the robot to further increase this probability. For example, the robot may be provided with an associated medication cache, carrier, magazine, or dispenser, and may also include with a water carrier or dispenser. When the compliance routine 54 is initiated by a time or other event trigger (for example, after a meal, upon direction from a remote caregiver, etc.), the robot may take the medication to the resident or bring the resident to the medication.
An exemplary regimen compliance routine, C, is depicted in
To increase probability of compliance, the robot may include a “snooze” function, SC9′, that permits the resident to temporarily delay the medication or regimen when the regimen is non-critical. The snooze routine may be limited in recurrences for particular regimens deemed more important, but not critical, by the resident, caregiver, or robot (e.g., medication compliance, therapeutic compliance, etc.). For less critical regimens (e.g., social and entertainment events, morning wake-ups on days with no events scheduled, etc.) the snooze routine may time out entirely. The snooze routine may also be modifiable with a secondary reminder (e.g., when permitting a delay by, for example, 15 minutes, the robot may issue a reminder that the compliance with the regimen will be soon required), and then return to SC3. The snooze function SC9′ may also be utilized when circumstances for compliance are unfavorable (e.g., the person is preparing a meal when a teleconference is scheduled, etc.). When a regimen compliance routine is interrupted or delayed by necessity and by the resident's control, it may be reinstated after a requested delay. Such delays may be managed by postponement rules that provide guidelines as to how many times and how long medication may be delayed or refused, and whether and how a caregiver is to be notified upon a delay or refusal.
If the routine is critical, the robot determines if the compliance counter has met a predetermined threshold, SC10, if not, the robot repeats the initial reminder at SC3. In the event the threshold is exceeded (resulting from persistent refusal by the resident), the robot is provided with more assertive reminders, SC11, allowing the robot to utilize other speech (“Bob, it is very important that you take this medicine”), or affect responses (Angry, Sad) to convince the resident to comply with the regimen. If the increased assertiveness reminder is not refused, SC12, the robot moves through the assistance subroutine, including steps SC13, SC14, SC15, and SC15′ (similar to the assistance subroutine beginning with step SC5 above) as needed. Continued refusal SC12 increases the compliance counter, SC16, until a second threshold is reached, at step SC17. As an additional compliance step, the robot may report continued refusal to the caregiver, with or without the resident's permission, depending on the application, SC18. Again, regimens deemed critical may be reported to the caregiver without resident permission, less critical regimens may require the resident's permission, or may go entirely unreported. Additionally, at the caregiver's or resident's discretion, the report to the caregiver may provide an opportunity to exercise caregiver intervention, e.g., including a videoconferencing call, where the caregiver may encourage the resident to follow the regimen. Alternatively, for caregivers without access to a videoconference system, the display of the robot may act as an avatar for the caregiver, wherein speech received from the caregiver is converted into viseme (and, optionally, phoneme) sequences on the display of the robot, to communicate with the resident and convince compliance.
The robot is also able to take other action in the event it is deliberately or inadvertently ignored, so that probability of compliance may be increased. Reasons for non-compliance can include simple or condition-dependent forgetfulness about what actions to take, or how those actions should be taken. Moreover, unavailability of medication at the designated time, other time or scheduling problems, the resident's inability to comply because of a medical condition, lack of understanding of the necessity of following the complete compliance regimen, and dislike of side effects or other disincentives may impact compliance. The robot may provide an option to call the caregiver to provide assistance when a compliance regimen is not followed or if other problems exist or arise at any point during the sequence. The caregiver may also remotely interrupt or conclude the performance of any interaction.
The camera and microphone on the mobile robot also allows the robot to operate as a video/teleconference station. As a videoconference station, the display may function as an avatar for the other party (as described above with regard to the caregiver avatar), or the robot may include a video coprocessor capable of processing high-compression video such as MPEG-4, H.264 or AVC, or other video and sound files.
The mobile robot may receive a conferencing request at step SV1 (for example, over a computer network using a wireless networking protocol, such as any standard IEEE 801.11 or 802.15, or BlueTooth, UWB or other impulse radio). The robot would then proceed to find or locate the resident at step SV2. Navigation sequences for resident searching and/or locating are described in more detail below. When the mobile robot finds the resident, it may issue an audible and/or visible query at step SV3, for example, “Shall I accept this conferencing request?”. At step SV4, the mobile robot may monitor the microphone, camera 28, and/or other sensors for verbal, gestural, or other responses from the resident. Additionally, the resident may touch an “Accept” button on one of the robot's screens, press a button on the surface of the robot, or touch a particular sensor. At step SV5, the robot determines from any one or more of the above events whether compliance was granted to begin the conferencing session. If permission was granted, the mobile robot may begin the conferencing session at step SV6; otherwise, the mobile robot may proceed to decline the session (and possibly, send notice of the decline to the remote conferencing requester) at step SV7.
To function as a useful companion to a resident, a robot may be able to navigate within an unpredictable home environment.
In constructing a topological map of the environment, the robot may move about the home and note rooms or other subdivisions of the environment. These subdivisions may be any logical division of a space, e.g., a hallway, quadrants or other portions of a large space, etc. The robot may identify the spaces at boundaries of rooms via “lighthouses,” or by visual identification of features (windows, etc.). Such space boundary identifiers are described in U.S. Provisional Patent Application Ser. No. 60/741,442, the disclosure of which is hereby incorporated by reference herein in its entirety. Additionally, the robot may recognize fiducials in the room, such as indicia projected onto the ceiling, self-similar tags, etc. Examples of fiducials that are projected onto a ceiling are described in U.S. patent application Ser. No. 11/176,048, the disclosure of which is hereby incorporated by reference herein in its entirety. The robot may also utilize position information, such as provided by odometry, optical or RF time of flight or angle of arrival triangulation or trilateralizing, or may use spatial estimates, e.g., room-size detection, size of largest diagonal, etc., to identify rooms. As the robot explores the various rooms/spaces within an environment, it may build a topological map to capture room adjacency information and maintain this information during normal operation. This map information is cached in the user/resident interface, or may be stored at an online site. The robot may also identify persons and things (a charging station, doors, windows, keys, pets, etc.), optionally by recognizing a radio frequency (RF) tag carried on, present on, or worn by the person or thing. Additionally, the robot could identify a person by a signature heat, voice, and/or visual recognition pattern, as described herein.
Once the robot completes, or as the robot completes, all or a portion of the topology (and resident/entity information), it may keep this information local or send this information to a web server, as previously discussed, (or serves the content itself). Discussed herein are methods for building an interface, an icon map, and meaningful connections between the map and the home, as well as methods for robot navigation using the combination of icons and room identity information. A web client may render an icon map of the home, which may in some embodiments appear as a collection of rooms arranged by distance on the topological map, not real-world distance. The web client may also draw and label figures or other markers for the resident or other entities within the house. Thus, the location of any identified person or thing within the house may be tracked. Such technology allows the resident or caregiver to monitor the location of objects, persons, or the like within the home, if these objects or persons are detectable to the robot (via tags or sufficient recognition resources). The resident or other user may rename or classify rooms: “this is a bathroom,” “this is my bedroom,” etc.
The location and identity of each of these rooms and entities may be stored within the robot memory or local or remote database, and presented to the resident upon request, for example, on a robot display or a remote caregiver station or web client. The robot may display a basic list of rooms and identified persons, things, etc., but preferably arranges the icon map as discussed herein. If desired, the resident may touch any room indicated on the screen to direct the robot to go to that room. As discussed herein, a privacy mode may require that the cameras of the robot are inaccessible to a remote caregiver when the robot is autonomously navigating from room to room, and require one-time or trust-level permission of the resident to be activated once the robot arrives in a room of interest. Once in an identified room (if permissions are appropriate), the robot may take a video or still image of the space, and either return to the resident to display this data, or send it remotely to the resident's computer, networked television, or personal digital assistant. The robot may also be used as a find tool, for keys or remote controls or the like. If equipped with suitable object recognition capability (e.g., scale invariant feature transform based object recognition and appropriate learning and database routines), the robot may maintain a list of people or things that it is currently tracking within the environment (e.g., according to where they were observed last). By selecting one of those items, the resident may direct the robot to proceed to the location where the particular person or thing is located.
As shown in
At step TM4, the robot captures topological adjacency of different rooms, and docks and chargers. Docks may charge and/or provide a data connection (e.g., such as a robot base station); chargers are intended to recharge the robot but may also have data access. “Capture” in this sense may include taking the entry of data from a person or floorplan analysis, but generally means closing (topological) paths and loops according to odometry or localization data, by routine or with the assistance of a person. At step TM 6, nodes are assigned by the robot. Simplicity means fewer nodes, e.g., one per room. However, bases and chargers may also have a node each, and hallways typically have a node whether or not they appear on an icon map. In addition, as shown in step TM6, doorways and interesting positions (e.g., facing to the TV, behind the chair, out of the way by the window) may also be assigned nodes. At step TM8, the number of rooms (e.g., the number of rooms a resident or caregiver has entered into the robot as existing in the house) are compared to the number of room identities (e.g., the number of rooms discovered by the robot or otherwise recorded by or made known to the robot). If there are too many room identities, the room identities are combined until the number of rooms equals the number of room identities.)
Once the topological map and/or number of rooms are in the robot or system, the user map or icon map can be built. The user map and topological map are distinct. The user map is built using a user interface presentation at a local or remote client (PC, dedicated, cell phone screen) or on the display of the robot. A person's identification of room identities and association with markers (or icons) provides the easiest and most reliable way of unifying the robot's and person's world views. The resident or caregiver may or may not be shown the contents of the topological map, and connectivity and adjacency from the topological map may or may not be used to display the user map (or be separately displayed superimposed or together with the user map).
As shown in
At step UM8, the robot (or web server, etc.) requests assignment of one of the room identity markers to the robot perspective or the current room (“Please confirm which room the robot is in?” from
Step UM14 is cumulative in nature, i.e., in order to add each assigned room identity to the icon map and complete the map, steps UM2-UM12 would be conducted for each iteration of step UM 14, until step UM 16 permits the process to conclude.
Directions via the icon map can provide multi mode semi-autonomous navigation, with decreasing granularity. First, the user selects a room using the room identity marker, and the robot uses the room identity and/or topological information to search for a room with the room identity until it is located (if a topological map is used, planning a route rather than searching).
In step IN-4, the robot may recognizes the present room. A variety of techniques are available as discussed herein, including localization via RF, fiducials, odometry, dead reckoning, object, pattern, surface, or feature recognition, and the like. If the room sought is not the one the robot is in (which can be a choice prevented by the interface), then the robot will begin to search for the room at step IN-6. The robot may follow a topological map, or may wall-follow, or may use SLAM or other navigation to begin this search. The search may also be essentially random. Once in a new room, the robot may stop at a predefined node, pose or arbitrarily, in a position that permits or facilitates room recognition. However, it is necessary that the robot identify the sought room when found—that a room the robot eventually finds corresponds to the sought room identity. Methods of recognizing a room are also discussed herein, and if the robot cannot confirm that it is in the correct room at step IN-8, the robot moves on to another room.
Webdrive mode has two sub-modes. As shown in
From a new position, although
As shown in
This navigation ability may also be remotely available to a resident who is away from home, or who may be confined to a particular area of the home. In such a case, the resident may remotely access the robot interface, via an internet connection (or via a personal digital device), and select a room or thing on which the resident desires status information. For example, the resident may want to robot to confirm that the front door is locked, and selects the living room on the map displayed on his or her web browser. The robot first identifies its present location (for example, the bedroom D). The robot then moves from the bedroom D, through the hallway F, until it reaches the living room B, where the front door is located. The robot may then take a visual image of the front door, and send that visual data to the remote resident, at which time the resident may instruct the robot to return to it's starting point, or to move to a different location. Alternatively, the resident may first select the room where it wants the robot to navigate to, then direct the robot to an item or more specific location within that room, then direct the robot to move on to a different location. The robot may also send streaming video to the resident as it moves throughout the home. Additionally, this mapping and tracking functionally may be available to a remote caregiver, with the appropriate permissions, as described herein.
In the course of interacting with a human resident, the robot and resident may become separated within an environment. Should the robot require contact with the resident, due to the beginning of a compliance routine, initiating a teleconference, or otherwise interacting, the robot will be able to navigate the environment to locate the resident. A plan view of one such navigation path in an environment 400 is depicted in
In each room, the robot 10 may employ a sensor suite that is responsive to the presence of a person (including sound detection, movement detection, pulse, breathing and/or heat detection) and/or an audible or visual query. The robot 10 also may move to a central area of the robot's known environment 400, or may move to an observation area in its present location (a central point in the room, for example), to begin its query. This first initial centralized locating function may be helpful if the robot recently had contact with the resident in the room or area in which it is still presently located but did not detect the resident leaving the room. The robot may also make a first round of known rooms, stopping in positions permitting surveying of large areas and monitor sensors for any responses. Filtering can be used to exclude false positives (the stove is too hot to be the resident, the dog is too low to be the resident, etc.).
In an alternative embodiment, the robot may utilize a paging system, with or without the use of sensors, to locate a person. This type of system may also be utilized if the robot is experiencing a sensor or other failure condition. When paging a person, the robot may use the doorways to separate rooms or chambers, in order to page the resident over only a short (in one embodiment, less than about 5 m) distance, from locations in which a direct line from robot to resident is more likely. The robot may also use this paging in subparts (smaller areas) of rooms or from passage outside of or between rooms. The robot may page from the doorways themselves in order to reach two rooms with one page.
When a person is detected, the robot may attempt to identify the person by employing an audible or visible query that asks whether the person is present, via a speaker and/or recorded speech or sounds, or synthesized speech or sounds. If the robot proceeds through every room, as depicted in
The mobile robot may also use a secondary system for confirming the identity of a person, such as by analyzing an infrared image or signature heat pattern corresponding to the person being sought, performing acoustic analysis to recognize the voice of the person, detecting an RFID or magnetic tag associated with the person or carried by the person, detecting a particular motion or a gesture, and/or performing image analysis on a video stream or still image frame generated by a camera to recognize facial features or memorized clothing sets typically worn by a person.
In selecting a next location to continue a search, the mobile robot may use any suitable method, including, but not limited to one or more of optical room feature recognition, sonar, RFID room tagging, infrared (IR) directional beacon detection, odometry, inertial guidance, dead reckoning, room mapping, and/or compass navigation. In one embodiment, the robot may use an available heading to a doorway or other wall gap (the heading made available by mapping, by passive detection of beacons from which the heading to a doorway can be inferred, or by active sensor detection of the doorway itself). The mobile robot may also choose an arbitrary path intended to locate a door in a reasonable time (e.g., wall following utilizing signal reflection) and then proceed until a doorway is detected as having been traversed. The robot may also identify a doorway by analyzing odometry and heading while in wall following mode. Identifying successive odometry/heading combinations may indicate that a doorway has been traversed. If the mobile robot encounters any obstacles when traveling to the next location, the mobile robot may circumnavigate the obstacle, or otherwise adjust its path of travel, in order to proceed to the coordinate location of the selected next location, or, alternatively, the mobile robot may simply halt at the obstacle and begin the query process anew. In the case of the robot adjusting its path of travel, once the robot passes an obstacle and encounters its previous heading, it may continue on until a predetermined distance is reached. In addition, the robot may simply move a threshold distance, ending in either another room, or in the same room, before beginning its query.
In addition, as illustrated in
When the mobile robot encounters a closed door, it may attempt to communicate with a person on the other side by applying an audible signal to the closed door. For example, the mobile robot may include a door knocker (such as a solenoid) or a speaker which can be applied to the closed door to transmit an audible signal or announcement to the closed door. Accordingly, the mobile robot can attempt to contact a person even when the person may be located in an area that is inaccessible to the robot.
Also, the mobile robot may plan a path to the selected next location based on its active or passive detections or map of the environment 400, or it may employ odometry, and/or inertial guidance when navigating. In this way, the mobile robot may identify when a door has been traversed, for example. Alternatively, the mobile robot may employ any suitable method for navigation, mapping, and room identification, such as, for example, infrared beam-detection (e.g., the navigation and beacon systems as disclosed in U.S. Provisional Patent Application Ser. No. 60/741,442, RFID room or door tagging, inductive or resonance detection of tank circuit or amorphous metal tags, optical feature detection or image analysis. The robot may also utilize Simultaneous Localization and Mapping (SLAM), using machine vision-based methods, computational resources, and sensors, to prepare vision and odometry data of an environment. Such systems are disclosed in U.S. Provisional Patent Application Ser. No. 60/822,636, the disclosure of which is hereby incorporated by reference herein in its entirety.
A locating routine depicted in
The robot's verbal annunciation need not sound like “shouting” or include blasting a claxon or other loud sound. Instead, the annunciator sound levels may be generally inoffensive to others in a room or adjacent room. As an example, normal background noise is about 35 dB, and the annunciator would sound in a normal speaking voice volume of about 55 dB to about 65 dB, or at least about 15 dB greater than detected background noise. If the resident has been identified as hard-of-hearing, the robot may increase the volume to up to about 75 dB or some other threshold level, and may add a visible signal and/or activate a resident-carried remote vibrator. The annunciator volume may be reduced to about 10 dB to about 20 dB greater than detected background noise; and/or may be pitched in frequencies that are typically still audible even to the elderly or those will various types of hearing loss (i.e., within a range of about 500 Hz to about 5000 Hz).
To facilitate a more rapid response to needs of the resident, or to limit the amount of time the robot spends away from the resident (thus encouraging more bonding), the robot may be provided with a sensor suite and reactive behaviors that tend to keep the robot in the vicinity of the resident, or track the resident from room-to-room. Simply following the resident from room to room may produce annoyance on the part of the resident to the constant presence of the robot; therefore, the robot instead undertakes actions and tasks that tend to keep the robot near to the resident, but without close following. The behaviors may (i) increase the probability that the robot is near to the resident, (ii) increase the probability that the robot is in a room or near to a room where the resident has been or will be, and/or (iii) increase the probability that the time to find the resident will be shorter.
For this purpose, in one embodiment, the robot may maintain a comfort or presence score (or anxiety score if the implementation is viewed from another perspective), which may increase, increment, or fully replenish in the resident's presence and may decrement, decay, or be reduced in jumps or steps as time passes or events occur without observing the resident. Once the comfort score decays beyond a threshold value or enters a certain range or region, the robot may register the score as a discomfort, which may activate or change the priority of a resident-seeking behavior or other behavior that tends to improve the proximity of the robot to the resident. Upon direct querying, sensing the resident, or checking phenomena that tend to correlate with the presence of the resident (heat, sound, movement), the comfort score may be replenished, or be gradually incremented, depending upon the character of the detection of the resident or phenomena.
An exemplary tracking routine that utilizes a comfort or presence score (as that score relates to proximity to the resident) is shown in
The robot would then park itself quietly, or with a simple comment or noise, in the same room as the resident, in a position that does not block traffic paths. Corridors and very small rooms such as bathrooms may be off-limits, and the robot may instead park itself in a nearby room. At this point, the cycle begins again as soon as the comfort score reaches a maximum. If the robot does not find the resident, it may seek the security of the charging station or other comfort generating phenomena. In this manner, although the robot does not follow the resident, the distance to the resident may usually not be very far, and the robot may be often in the same room. The robot may also communicate in order to elicit a response from the resident and/or provoke interaction. Accordingly, the robot may be perceived as a more natural entity, may maintain proximity to the resident without uninvited irritating following behaviors, and enhance the likelihood of successfully interaction with the person, as well as ensuring that the resident is quickly locatable. By using the tracking behavior, the robot increases the probability that it will be near to the resident when the resident is to be found or when the robot is called by the resident. The robot need not be within detection range, only closer to the resident than it was when it started looking, although parking within a detection range may require that the robot search less often. If a charging station is nearby, the robot dock and charge. Additionally, the robot may dock and charge with a nearby charger, even if that charger lies outside of the detection range, if the robot concludes it requires additional power.
Alternatively or in addition, the robot may behave in accordance with a virtual pheromone system, in which a room is assigned the comfort score rather than one score kept by the robot. When the resident is not observed in a particular room or location, on the other hand, the comfort score of that room would decline. Entering a room of frequent use may increment the comfort score, and the robot may tend to navigate to rooms that tend to increase the comfort score. These properties may be time dependent—the robot may look first in the TV room at 7 pm, first in the kitchen at noon, and first in the bedroom after bedtime; the decay rate may be correlated with times of frequent room changing (early morning activity) or stability (evening leisure activities). Thus, the robot may choose to seek out the resident in rooms with higher comfort scores before resorting to searching locations with lower comfort scores, and therefore potentially increase the likelihood and typical speed of looking for the person. Alternatively or in addition, the robot may form a comfort-related spatial occupancy map, landmark map, or topological map that updates features or cells according to the probability that the feature or cell is adjacent to or occupied by the resident—in this case, each cell corresponding to a room, corridor, or a nearly room-sized area.
Alternatively or in addition, the robot may behave in accordance with a force, vector or potential field that maps a comfort scores/vectors onto a motor, drive, or navigation goal space. The comfort scores may be vectors or areas of high comfort potential, which may decay over time. The comfort potential of the resident in the potential field may be compared to with the comfort or security potential of the charger, base station, or recorded wireless signal strength field to decide the location where the robot will park itself in the room with the resident. Such fields can define that the robot should be near to, but not too near to, the resident, and also in a strong wireless signal strength area available in that room.
The mobile robot may improve its rate of successfully maintaining proximity to the resident by tracking and following the resident in certain circumstances: e.g., a fixed or variable percentage of the time the robot successfully detects that the resident has left the room, and/or during certain times of the day, and/or after certain patterns of conduct by the resident. For example, if the robot recognizes that the resident is beginning preparations to retire for the evening, the robot may follow the resident from living room to bedroom. Additionally, or in alternative, the robot may record, interpret, or accept inputs describing the resident's daily routine or schedule, and may use the schedule to predict and move to an appropriate room ahead of the resident.
The robot may initially become aware of the resident's presence by any appropriate method, such as by asking whether the resident is present and receiving an affirmative response, or by having the robot's attention directed to the resident via a command or input from the resident or a caregiver, for example. The robot may detect the presence of the resident using a physical sensor or input from a variety of sensors, such as infrared sensors, heat sensors, motion sensors, optical sensors, a microphone or other acoustic sensor, electrical property sensors (e.g., capacitance) or any other suitable physical sensor. Alternatively, the resident may carry a tag such as an active or passive RF or RFID tag, tank circuit, sonic emitter, light source, any of which may carry a modulated signal, or the like, which the robot may use to identify and locate the resident. Used in conjunction with an item such as a medical alert bracelet would make such an item very useful for increasing the reaction time of the robot to various medical situations. As another example, once the robot has located the resident in order to perform a task (as described in the above-described embodiments), the robot may then continue to keep track of the location of the resident and track the resident as the resident moves about the environment.
While navigating an environment, the robot will encounter various candidate objects (objects higher than room temperature, in motion, or generating noise), which may or may not be the resident sought. In order to determine whether a candidate object has a high likelihood of being the resident, the robot may associate a particular sensor input profile with the resident, and compare any objects that come within range of the robot with the resident's sensor profile. For example, if Ann is the resident that is being cared for, then the robot may record Ann's average temperature using an IR sensor, Ann's approximate size using a camera and/or image analyzer, the clothing combinations typically worn by Ann in the same manner, pitch of Ann's voice or the sound of Ann's breathing using a microphone or acoustic sensor, etc., during the course of performing tasks with or for Ann. Based on these or other inputs, the robot may generate a sensor profile that is characteristic of Ann. Furthermore, the robot may analyze or distill the characteristic sensor profile of the resident by employing suitable statistical models or recognition models, e.g., Bayesian analysis, Hidden Markov Models, techniques dependent upon these two, or other statistical or heuristic routines for filtering and recognizing.
The robot may also track a number of grouped characteristics of an entity, maintain a persistent record of grouped characteristics labeled as a particular entity, and/or the location of that entity in space (including topological space or other abstract representative space). When the robot compares an observed object to the resident's characteristic sensor profile, it may generate a comparison score that generally correlates to the likelihood that the observed object is, in fact, the resident. If the comparison score of the observed object exceeds a threshold score, the robot may track the object in order to remain generally proximal to the resident, or may rank the comparison scores of two or more of the observed objects to determine which object is more likely to be the resident, or may track new objects instead of old objects, or may switch tracking from one object to another when the comparison yields a difference above threshold. The robot may rely on its likelihood analysis without seeking direct identity confirmation from the resident. The robot may try to maintain a particular distance or distance range from the resident, and may react only after a delay or distance threshold is crossed.
Within an environment, one or more docks or base stations (such as a recharging dock or resupply dock) may be positioned for the robot to recharge its batteries or obtain supplies. A location of one such charging station is depicted in
A regimen compliance assistance routine is detailed in
The robot checks the information often enough to anticipate that it will need to be near the resident at a certain time in the future (e.g., frequently checking ahead to accommodate changes in schedule, or by setting ticklers ahead of the event). When an event is upcoming (anticipated or known due to a tickler/reminder) at RC-4, the robot begins to look for the person/resident. In the case of the proximity maintaining routines discussed herein, the likelihood that the resident is very near is good. However, even absent the proximity maintaining routines, the robot need only schedule enough time in advance to traverse the entire household, and in most cases will not need to. Rather than employ proximity events as previously described, the robot may learn a schedule by recording the rooms the resident is found in every day and time an event is needed, and in short order will learn the most likely room, the second most likely room, etc., for a time of day. The person finding routine employed at step RC-8 is as described herein, and may include a verification step to check the identity of the person.
When the resident has been located and their identity verified, the robot establishes a position next to the resident. The resident may be otherwise engaged, and so the robot takes up a position that is preferably unobtrusive (E.g., against a wall, in a nearby charging dock, in a corner of the room, beside the arm of the chair in which a resident sits) and awaits the scheduled event. In the case of a generalized regimen compliance event, at RC-16 the resident is reminded of the regimen activity they are expected to accommodate. For exercise, the robot may act as coach, may display exercises on screen, may link into a live exercise session, and may synchronize its own motions with those exercises to be performed by the resident (along with other appropriate movement expressions, modulated and controlled as discussed herein to match particular scripted events or non-scripted stimuli). For therapy, the robot may display therapies on screen, may guide the resident to particular equipment (or carry the same), may link into a live or recorded therapist, and may take the resident's reports of status information (e.g. following exercise or therapy, reporting heart rate, blood pressure, breathing rate, either by having the resident measure them or by using on-board interactive sensors).
It should be noted that as described herein, the robot contains full connectivity and support for power, network access, security, and mobility to health assisting appliances, one example of which is a “Health Buddy” available from Health Hero Network, Inc., Redwood City, Calif., and such as those discussed in U.S. Pat. No. 5,307,263, herein incorporated by reference in its entirety. A typical device is a 1.25 lb device with a small color LCD screen, soft buttons, and medical instrument and network interfaces. It includes may include USB ports, an RJ-45 Ethernet network port, and embedded-class memory (64 MB) and processing (ARM processor, Linux OS). The devices are typically powered by 5-24V DC via an AC adapter. The contemplated robot can provide more robust, longer lasting power (2400-9000 milliamp-hours) for appliance or instruments, as well as smart control of on-off cycles (turning on the Health Buddy only as needed); mobility and patient finding for the appliance itself as well as instruments, additional network connectivity to additional instruments, wireless devices, pagers, telecommunications; full-screen, full-support, and live tutorials and instructions; video conferencing, remote observation, statistical reporting, richer controls and interfaces. Such appliances may be connected via network interface to collect data from to blood glucose meters, peak flow meters, scales, blood pressure monitors, pulse oximeters and the like, and robot may intercept, split, transfer or otherwise link to the same instruments, and the regimen compliance reminder may hand off to or receive delegation from a Health Buddy or like device pursuant to independent or linked schedules, and subsequently route collected data, interface events, communications, and all other digitally transmissible information discussed herein via the appliance's supporting networks or the robot networks discussed herein. The Health Buddy is a Linux device having a limited subset of the connectivity and user interface elements of a robot as discussed herein, and all of the Health Buddy processes, software and interfaces, including user interfaces displayable on a screen and controllable by user interface elements such as a touch screen or soft buttons, may be directly run by the robot discussed herein. As discussed herein, regimen compliance, scheduling, interaction, reporting and all other compliance and medical reporting discussed herein, as supported by resident or patient finding, reminding, household navigation and resident location awareness, medication loading and carrying, privacy and permission administration, and human-robot interaction, expressly includes all uses of appliances such as the Health Buddy supported by, carried by, and or linked to the mobile robot described herein, and the claims herein expressly consider such an accessory appliance to be fall within the literal language of a “robot,” “robot system,” or method of operating, navigating, or otherwise controlling a robot or robot system.
If the compliance event is associated with the taking of medications, steps RC-12 and RC-14 are specifically applicable. In the context of these claims, “matching” is the task of putting the resident in the same place at the same time as their medication, and may involve bringing the medication to the person or guiding the person to the medication. In a failure mode where mobility is not working, matching is replaced by mere reminders. In step RC-12, a human perceptible signal (recorded speech, synthesized speech, a displayed picture or displayed text, flashing or blinking lights, pointers, outlines, or highlights) is “sent” to the resident, and shows the resident the location where the medication is to be found (which may in fact be inches away carried by the robot, or may be in a room where a medication dispenser that is a networked part of the robot system is kept, or may be where medication is normally kept for that resident, as well as other variations discussed herein). Guiding, as discussed in step RC-14, is the task of showing or leading the resident to the location of the medication, and may be showing an image of the medication on the robot, a pointer or indicator drawing attention to a robot-carried dispenser, or actually moving the robot toward the bathroom or other room where a dispenser is located while expressing that the resident should accompany the robot by spoken messages, sounds, or expression motions.
One kind of matching is shown in
In step ML-4, the resident is asked whether or not he or she is ready to load the dispenser. This is optionally handled as any other compliance event discussed herein, and can be subject to snooze delays, reporting to the caregiver, and like procedures detailed herein. A simple dialogue is shown at MSG-4 in which the robot checks whether the resident has left-over medication in the dispenser, and/or expects that he or she has all medication available. In step ML-6, the robot repeats and steps through each medication. A simple dialogue is shown at MSG-6, in which the robot explains the particular regimen for a medication, uses the name of the medication, shows an image of the medication, and requests the user to load a chamber, compartment, disposable cup, or other holder with the medication. More medications are stepped through in the same way. The routine does not show more significant interaction (e.g., such as the resident wishing to ask about side effects or other queries), but more responsive dialogues (such as are discussed herein) would permit this to occur. Once the medications have all been loaded, the robot may ask the resident to quickly step through and double-check the day's complement (ML-8). If the caregiver is permitted to or designated to check the medication complement (ML-10), the robot may open a communication channel to the caregiver (step ML-10, the same as or similar to the above discussion of steps RC-18 and step RC-20), and then permit the caregiver to ask questions of the resident (or for a richly instrumented dispenser, directly check the status). The dispenser may be arranged (with a transparent or open cover) and positioned so that it is visible to a camera of the robot, and indeed the resident may hold a dispenser so configured up to one of the robot's cameras.
In addition to the above-identified functionality, the robot may be equipped with RFID or barcode sensors for assisting with shopping, etc. In an example, the robot may accompany the resident to a store or supermarket, in which the items for sale include RFID tags (or barcodes, or any other suitable identification scheme) for identification and inventory. If the resident wants to purchase an item, the robot may scan the RFID information of the indicated product, and relay the information to the store for purchasing assistance. In addition, the robot may keep track of food or supply inventories in the household by scanning RFID tags or barcodes of items in the refrigerator or pantry. When items run low, the robot may retrieve additional supplies and restock, for example. Such functionality is not limited to only food or store items, but may also be applied to any suitable items, such as clothing, automobile parts, etc.
In addition, the robot may assist in keeping the household clean, by monitoring clutter or by keeping track of cleaning schedules. A clutter level can be scored according to, e.g., (i) a number of collisions or proximity detections per unit time or per unit distance traveled (these could be separated into wall, furniture, and non-wall detections based on sensor data); and/or (ii) a number of differently colored or differently textured items per unit area detected by cameras viewing the ground plane in front of the robot; and/or (iii) a number of medium-range obstacles detected per unit time, distance, or area by ranging sensors. In this manner, the robot may apply image analysis to views of the house obtained from an on-board camera, and determine if the house has become too messy, or may determine that clutter has accrued when the robot too frequently encounters obstacles along paths that should be or are historically clear, for example.
The robot may facilitate bill paying by sharing checklists between the resident and a remote caregiver. Accordingly, when a bill payment or other household task is due, the robot may present a checklist of tasks to the resident on its display screen (and may also verbally or audibly inform the person). If tasks have not been completed, the robot may notify a caregiver, if concordant with the privacy permissions. Also, the task checklists may be programmed by either the resident or by a remote caregiver, for example. The robot's cameras also may be used to photograph and/or OCR bills and notices.
If the resident requires assistance dealing with an outside third party or service provider, such as a housing contractor, salesperson, or nurse or aide, the robot may provide on-the-spot teleconferencing between the people present at the house and the remote caregiver, for example. Furthermore, by providing remote visual inspection capabilities, the robot may facilitate housing safety inspections by permitting a remote inspector (or caregiver) to view various locations of the house. The robot may also include safety sensors such as a smoke detector, radon gas detector, carbon monoxide or CO2 sensor, or the like.
The robot may use an infrared emitter to control one or more home entertainment devices, such as televisions, VCRs, DVD players, or stereo systems. For example, the robot may function as a device coordinator, and control several pieces of equipment simultaneously or according to a sequence, such that the robot could turn on both the television and DVD player (and switch a stereo mixer to input from the DVD player, for example) when the resident wants to view a DVD. As another feature, the robot may be easier to locate than a small remote control, and the robot may also include a database of remote control codes for various products and manufacturers, relieving the resident of having to program a remote control for multiple items.
Furthermore, the robot (or other appliance having a camera and IR modulated emitter) may detect RFID, barcode, or brand trademarks on the home equipment, and discern the appropriate remote control codes therefore without requiring further input from the person. The robot (or remote control or other appliance, having a connection to a camera and IR modulated emitter) may photograph part of or the entire front face panel of any remote controlled equipment (TV, stereo, video recorder, cable box, CD/DVD/disc player, home theater, amplifier, tuner, lighting, fans, air conditioners, video games, etc.). The robot may then relay an image of the equipment to a remote computer, which may analyze the image for brand information or model information, and relay the appropriate remote control code back to the robot, for example, and therefore the robot need not necessarily include image processing capabilities locally.
Visual pattern recognition (which may employ a scale invariant feature transform “SIFT” system such as the ViPR™ system, Evolution Robotics of Pasadena, Calif., or other known scale invariant feature transform algorithms such as “SURF”) may identify features or distinct regions and compute from 1-1000 descriptors of unique visual patterns, which in combination may be used to uniquely identify the brand and model (or model family) of equipment, including OEM variations. Any portion of this process can be carried out on the robot or remotely on a server; in most cases the robot would simply send the image to the server for decomposition and comparison to the database of trained equipment images. A voting or probabilistic routine can employ feature votes to cross-correlate features in trained images to similar features in the recorded image. The votes are tabulated to identify the most likely identity for the equipment or different items of equipment. The robot selects or receives the appropriate libraries of remote control IR modulated codes for use with every device photographed (e.g., all of the codes for each of the TV, DVD player, and cable box). A single photograph or several photographs may be decomposed or analyzed to identify every piece of equipment in one place, in a room or in the residence.
The robot can then employ a compatibility routine to control the devices according to the overall task specified by the resident, by coordinating compatibility of inputs and outputs (e.g., in response to a command to watch TV, the robot may activate, via the received IR codes: the TV ON, the DVD OFF, the cable box ON, and the TV to the CABLE input; in response to a command to watch a DVD, the robot may activate, via the received IR codes, the TV ON, the DVD ON, the cable box OFF, and the TV to the S-VIDEO input). In this manner, the robot would operate similarly to the family of Harmony™ Remote Controls sold by Logitech Inc., of Fremont, Calif., using a similar database of IR codes and products. The Harmony devices are programmed by the user, suffer from drift when one controlled device is no longer synchronized from the others, and employ an all-off routine to reset all devices to a default state in order to recover the input/state synchronization. In contrast, the present robot or remote control or other appliance, having a connection to a camera and IR-modulated emitter may use the camera and aforementioned visual pattern recognition to detect the on, off, or other active (e.g., wrong input selected, volume low) state of any device by status lights, status indications (a picture on the TV), status messages (on the TV screen or device display) or other pattern recognition to avoid using an all-off routine, instead resynchronizing the devices to one another based on the present state of inputs and other functions. The robot or other device may also use speech recognition to receive verbal remote control commands from the person, for example.
With regard to verbal input and/or output, the mobile robot may include a database of verbal patterns for various events and/or tasks, such that the robot may correctly interpret verbal input associated with a particular task. For example, when the robot is performing a medication-related task, the robot may load medical terminology into its immediate speech-recognition pattern memory; if the robot later performs a shopping task, the robot may switch to a clothing- or grocery-related speech pattern database.
Many examples are given herein that relate to elder care. However, the same embodiments of the invention are readily adaptable for care of children or in limited cases, adults who require caregiver assistance. In order to provide mental stimulation to the person, the robot may use sound, music, and/or games, either spontaneously, or in response to a request by the resident or caregiver, or in accordance with a predetermined schedule. In addition, the robot may perform any of the following interactive activities: play games using a touch screen; play word games (audio only, or via the display screen, or both); play music; play audio books; tell stories or recite audio books; play daily news stories, weather, or sports scores; look up reference information (encyclopedia, TV guides) (in response to verbal input or questions from the person, for example); look up financial information (stock prices, bank balance); receive emails and instant messages; show digital photos sent by family and friends; play videos sent by family and fiends; talk on the phone; participate in a video conference; produce a daily report sent to a caregiver in email that lists activities the resident performed that day; play audio clips of the resident talking to the robot; show photos of the resident in their home; give medication reminders including dosage and timing; give physical activity reminders (e.g., go for a walk, do stretches); give daily event reminders (e.g., TV show times, appointments, birthdays); let a caregiver talk to and see the resident from their computer; let a caregiver navigate the robot around the person's house to check on things; include privacy control (the robot requires permission from the resident before turning on a camera or letting a remote caregiver do a virtual visit; include a trip detector (the robot detects when the resident is nearby, and gets out of the way or makes sounds and flashes lights); include caregiver management; and/or give updates (new games, music, videos and photos may be sent to the robot automatically via an Internet feed, or specifically by a caregiver or third party from a remote location (via e-mail or web interface)).
The invention has been described in detail in connection with various embodiments. These embodiments, however, are merely for example only and the invention is not limited thereto. It will be appreciated by those skilled in the art that other variations and modifications can be easily made within the scope of the invention as defined by the appended claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US6542788 *||Dec 29, 2000||Apr 1, 2003||Sony Corporation||Robot apparatus capable of selecting transmission destination, and control method therefor|
|US6658325 *||Jan 14, 2002||Dec 2, 2003||Stephen Eliot Zweig||Mobile robotic with web server and digital radio links|
|US7388879 *||Aug 28, 2001||Jun 17, 2008||Sony Corporation||Communication device and communication method network system and robot apparatus|
|US7467026 *||Aug 13, 2004||Dec 16, 2008||Honda Motor Co. Ltd.||Autonomously moving robot management system|
|US7624438 *||Aug 19, 2004||Nov 24, 2009||Eric White||System and method for providing a secure connection between networked computers|
|US20040019406 *||Jul 25, 2002||Jan 29, 2004||Yulun Wang||Medical tele-robotic system|
|US20040078219 *||Oct 21, 2002||Apr 22, 2004||Kimberly-Clark Worldwide, Inc.||Healthcare networks with biosensors|
|US20040093650 *||Dec 4, 2001||May 13, 2004||Martins Goesta||Robot system|
|US20050091684||Sep 22, 2004||Apr 28, 2005||Shunichi Kawabata||Robot apparatus for supporting user's actions|
|US20050216126 *||Mar 26, 2005||Sep 29, 2005||Vision Robotics Corporation||Autonomous personal service robot|
|US20060052676 *||Sep 7, 2004||Mar 9, 2006||Yulun Wang||Tele-presence system that allows for remote monitoring/observation and review of a patient and their medical records|
|EP1548532A1 *||Oct 1, 2002||Jun 29, 2005||Fujitsu Limited||Robot|
|JP2004337556A||Title not available|
|1||"Alfred: The Robot Waiter Who Remembers You" Maxwell et al., XP003001345 Swarthmore College, Swarthmore, PA 19081, Maxwell@email@example.com.|
|2||*||D.B.Baker et al. "PCASSO: Applying and Extending State-of-the-Art Security in the Healthcare Domain" , 1997 ACSA conf. pp. 251-266, IEEE.|
|3||*||D.R.Masys et al. "Patient-Centered Access to Secure Systems Online (PCASSO): A Secure Approach to Clinical Data Access Via the World Wide Web", Proc pf 1997 AMIA Fall Symp. pp. 340-343.|
|4||International Search Report and Written Opinion for PCT/US2006/038063, dated May 7, 2007.|
|5||*||V.Stanford, Pervasive Health Care Applications Face Tough Security Challenges? Pervasive Computing, 2002 IEEE, pp. 8-12.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US8121728 *||Jun 6, 2008||Feb 21, 2012||Hong Fu Jin Precision Industry (Shen Zhen) Co., Ltd.||Robot apparatus and output control method thereof|
|US8265793 *||Sep 27, 2007||Sep 11, 2012||Irobot Corporation||Mobile robot for telecommunication|
|US8301305 *||Mar 22, 2010||Oct 30, 2012||Msi Computer (Shenzhen) Co., Ltd.||Moving devices and controlling methods therefor|
|US8359122 *||Feb 15, 2007||Jan 22, 2013||Vision Robotics Corporation||Autonomous personal service robot|
|US8396254 *||Aug 17, 2012||Mar 12, 2013||Google Inc.||Methods and systems for estimating a location of a robot|
|US8414349||Jun 1, 2011||Apr 9, 2013||Nintendo Co., Ltd.||Remotely controlled mobile device control system|
|US8417384 *||Jun 19, 2008||Apr 9, 2013||Sony Corporation||Information processing system, robot apparatus, and control method therefor|
|US8423225 *||Nov 11, 2009||Apr 16, 2013||Intellibot Robotics Llc||Methods and systems for movement of robotic device using video signal|
|US8433442||Jan 28, 2009||Apr 30, 2013||Seegrid Corporation||Methods for repurposing temporal-spatial information collected by service robots|
|US8532822 *||Oct 24, 2011||Sep 10, 2013||F Robotics Acquisitions Ltd.||System and method for determining the location of a machine|
|US8577501 *||Jul 31, 2012||Nov 5, 2013||Irobot Corporation||Mobile robot for telecommunication|
|US8588972||Apr 17, 2011||Nov 19, 2013||Hei Tao Fung||Method for creating low-cost interactive entertainment robots|
|US8660693 *||Dec 15, 2010||Feb 25, 2014||Electronics And Telecommunications Research Institute||Component integration apparatus and method for collaboration of heterogeneous robot|
|US8679260||Nov 11, 2009||Mar 25, 2014||Intellibot Robotics Llc||Methods and systems for movement of an automatic cleaning device using video signal|
|US8745634 *||Oct 15, 2010||Jun 3, 2014||Invensys Systems, Inc.||System and method for integrated workflow scaling|
|US8755936 *||Feb 13, 2009||Jun 17, 2014||Seegrid Corporation||Distributed multi-robot system|
|US8774950 *||Jan 22, 2009||Jul 8, 2014||Carnegie Mellon University||Apparatuses, systems, and methods for apparatus operation and remote sensing|
|US8838268||Jan 28, 2009||Sep 16, 2014||Seegrid Corporation||Service robot and method of operating same|
|US8862271 *||Sep 23, 2013||Oct 14, 2014||Irobot Corporation||Proximity sensing on mobile robots|
|US8868238 *||Jan 10, 2013||Oct 21, 2014||The United States Of America As Represented By The Secretary Of The Army||Apparatus and method for systematic control of robotic deployment and extraction|
|US8892256||Jan 28, 2009||Nov 18, 2014||Seegrid Corporation||Methods for real-time and near real-time interactions with robots that service a facility|
|US8892260 *||Sep 30, 2013||Nov 18, 2014||Irobot Corporation||Mobile robot for telecommunication|
|US8909370 *||May 8, 2008||Dec 9, 2014||Massachusetts Institute Of Technology||Interactive systems employing robotic companions|
|US8930023||Nov 5, 2010||Jan 6, 2015||Irobot Corporation||Localization by learning of wave-signal distributions|
|US8957774||Feb 22, 2011||Feb 17, 2015||Vivian B. Goldblatt||Concealed personal alarm and method|
|US8958912||Sep 17, 2012||Feb 17, 2015||Rethink Robotics, Inc.||Training and operating industrial robots|
|US8965576||Sep 17, 2012||Feb 24, 2015||Rethink Robotics, Inc.||User interfaces for robot training|
|US8965580||Sep 17, 2012||Feb 24, 2015||Rethink Robotics, Inc.||Training and operating industrial robots|
|US8972061||Mar 8, 2013||Mar 3, 2015||Irobot Corporation||Autonomous coverage robot|
|US8989077 *||Jun 10, 2010||Mar 24, 2015||Sony Corporation||Base station, communication system, mobile terminal, and relay device|
|US8996174 *||Sep 17, 2012||Mar 31, 2015||Rethink Robotics, Inc.||User interfaces for robot training|
|US8996175||Sep 17, 2012||Mar 31, 2015||Rethink Robotics, Inc.||Training and operating industrial robots|
|US9026302||Nov 5, 2010||May 5, 2015||Irobot Corporation||Methods and systems for complete coverage of a surface by an autonomous robot|
|US9044855 *||Jun 20, 2012||Jun 2, 2015||International Business Machines Corporation||Processing context information|
|US9044863||Feb 6, 2014||Jun 2, 2015||Steelcase Inc.||Polarized enhanced confidentiality in mobile camera applications|
|US9047052||Dec 22, 2009||Jun 2, 2015||At&T Intellectual Property I, L.P.||Simplified control input to a mobile device|
|US9092698||Sep 17, 2012||Jul 28, 2015||Rethink Robotics, Inc.||Vision-guided robots and methods of training them|
|US9182763 *||May 19, 2008||Nov 10, 2015||Samsung Electronics Co., Ltd.||Apparatus and method for generating three-dimensional map using structured light|
|US9188983||Apr 11, 2014||Nov 17, 2015||Irobot Corporation||Methods and systems for complete coverage of a surface by an autonomous robot|
|US9296109 *||Oct 13, 2014||Mar 29, 2016||Irobot Corporation||Mobile robot for telecommunication|
|US9310802 *||May 28, 2015||Apr 12, 2016||Jaybridge Robotics, Inc.||Multi-operator, multi-robot control system with automatic vehicle selection|
|US9408515||Dec 31, 2014||Aug 9, 2016||Irobot Corporation||Autonomous coverage robot|
|US9431027||Jan 17, 2012||Aug 30, 2016||Honda Motor Co., Ltd.||Synchronized gesture and speech production for humanoid robots using random numbers|
|US9434072||Sep 17, 2012||Sep 6, 2016||Rethink Robotics, Inc.||Vision-guided robots and methods of training them|
|US9436922 *||Apr 2, 2014||Sep 6, 2016||Invensys Systems, Inc.||System and method for integrated workflow scaling|
|US9505127 *||Dec 25, 2012||Nov 29, 2016||Sharp Kabushiki Kaisha||Self-propelled electronic device|
|US9547112||Apr 27, 2015||Jan 17, 2017||Steelcase Inc.||Polarized enhanced confidentiality|
|US9603499||Sep 16, 2014||Mar 28, 2017||Seegrid Corporation||Service robot and method of operating same|
|US9609991||Feb 10, 2014||Apr 4, 2017||Diversey, Inc.||Methods and systems for movement of an automatic cleaning device using video signal|
|US9669544||Jun 24, 2015||Jun 6, 2017||Rethink Robotics, Inc.||Vision-guided robots and methods of training them|
|US9701015||Jun 24, 2015||Jul 11, 2017||Rethink Robotics, Inc.||Vision-guided robots and methods of training them|
|US9704043||Dec 16, 2014||Jul 11, 2017||Irobot Corporation||Systems and methods for capturing images and annotating the captured images with information|
|US9729824 *||Jul 27, 2015||Aug 8, 2017||Microsoft Technology Licensing, Llc||Privacy camera|
|US9762717||Jun 1, 2015||Sep 12, 2017||At&T Intellectual Property I, L.P.||Simplified control input to a mobile device|
|US9811089||Dec 19, 2013||Nov 7, 2017||Aktiebolaget Electrolux||Robotic cleaning device with perimeter recording function|
|US20070198129 *||Feb 15, 2007||Aug 23, 2007||Harvey Koselka||Autonomous personal service robot|
|US20080215183 *||Feb 27, 2008||Sep 4, 2008||Ying-Tsai Chen||Interactive Entertainment Robot and Method of Controlling the Same|
|US20080306629 *||Jun 6, 2008||Dec 11, 2008||Hong Fu Jin Precision Industry (Shen Zhen) Co., Ltd.||Robot apparatus and output control method thereof|
|US20090033470 *||Dec 28, 2007||Feb 5, 2009||Industrial Technology Research Institute||Apparatus for robot behavior series control based on rfid technology|
|US20090037023 *||Jun 19, 2008||Feb 5, 2009||Sony Computer Entertainment Inc.||Information processing system, robot apparatus, and control method therefor|
|US20090055019 *||May 8, 2008||Feb 26, 2009||Massachusetts Institute Of Technology||Interactive systems employing robotic companions|
|US20090125175 *||May 19, 2008||May 14, 2009||Samsung Electronics Co., Ltd.||Apparatus and method for generating three-dimensional map using structured light|
|US20090198376 *||Feb 13, 2009||Aug 6, 2009||Seegrid Corporation||Distributed multi-robot system|
|US20090198381 *||Jan 28, 2009||Aug 6, 2009||Seegrid Corporation||Methods for repurposing temporal-spatial information collected by service robots|
|US20100076600 *||Sep 27, 2007||Mar 25, 2010||Irobot Corporation||Mobile robot for telecommunication|
|US20110066262 *||Jan 22, 2009||Mar 17, 2011||Carnegie Mellon University||Apparatuses, Systems, and Methods for Apparatus Operation and Remote Sensing|
|US20110108064 *||Nov 11, 2009||May 12, 2011||Intellibot Robotics, Llc||Methods and systems for movement of an automaic cleaning device using video signal|
|US20110112714 *||Nov 11, 2009||May 12, 2011||Intellibot Robotics, Llc||Methods and systems for movement of robotic device using video signal|
|US20110130874 *||Mar 22, 2010||Jun 2, 2011||Chuan-Ching Tsao||Moving devices and controlling methods therefor|
|US20110151929 *||Dec 22, 2009||Jun 23, 2011||At&T Intellectual Property I, L.P.||Simplified control input to a mobile device|
|US20110153077 *||Dec 15, 2010||Jun 23, 2011||Electronics And Telecommunications Research Institute||Component integration apparatus and method for collaboration of heterogeneous robot|
|US20110167574 *||Nov 5, 2010||Jul 14, 2011||Evolution Robotics, Inc.||Methods and systems for complete coverage of a surface by an autonomous robot|
|US20110246551 *||Jun 20, 2008||Oct 6, 2011||Space Software Italia S.P.A.||Adaptive multifunction mission system|
|US20120041594 *||Oct 24, 2011||Feb 16, 2012||F Robotics Acquisitions Ltd.||System and Method for Determining the Location of a Machine|
|US20120096463 *||Oct 15, 2010||Apr 19, 2012||Invensys Systems Inc.||System and Method for Integrated Workflow Scaling|
|US20120113888 *||Jun 10, 2010||May 10, 2012||Sony Corporation||Base station, communication system, mobile terminal, and relay device|
|US20120328092 *||Jun 20, 2012||Dec 27, 2012||International Business Machines Corporation||Processing context information|
|US20130035806 *||Jul 31, 2012||Feb 7, 2013||Matthew Cross||Mobile Robot for Telecommunication|
|US20130173212 *||Dec 10, 2012||Jul 4, 2013||Seiko Epson Corporation||Motion analysis method and motion analysis apparatus|
|US20140088761 *||Sep 23, 2013||Mar 27, 2014||Irobot Corporation||Proximity Sensing On Mobile Robots|
|US20140214478 *||Apr 2, 2014||Jul 31, 2014||Invensys Systems, Inc.||System and method for integrated workflow scaling|
|US20140218516 *||Jul 1, 2013||Aug 7, 2014||Electronics And Telecommunications Research Institute||Method and apparatus for recognizing human information|
|US20140379127 *||Dec 25, 2012||Dec 25, 2014||Sharp Kabushiki Kaisha||Self-propelled electronic device|
|US20150094854 *||Oct 13, 2014||Apr 2, 2015||Irobot Corporation||Mobile Robot for Telecommunication|
|US20150334348 *||Jul 27, 2015||Nov 19, 2015||Microsoft Technology Licensing, Llc||Privacy camera|
|USD765180 *||Apr 9, 2015||Aug 30, 2016||The Coca-Cola Company||Dispenser robot with animated facial expressions|
|USD766644 *||Mar 5, 2015||Sep 20, 2016||The Coca-Cola Company||Dispenser robot|
|USD793145 *||Aug 11, 2016||Aug 1, 2017||The Coca-Cola Company||Beverage container dispenser robot|
|CN104982029A *||Dec 19, 2013||Oct 14, 2015||微软技术许可有限责任公司||CAmera With Privacy Modes|
|U.S. Classification||700/245, 700/249, 340/3.5, 700/264, 370/351, 709/223, 709/229, 713/153|
|International Classification||G06F19/00, H04W48/00|
|Cooperative Classification||B25J9/0003, B25J11/008, G06F19/3462, Y10S901/46, G06F19/3406, G05D2201/0207, G05D1/0242, G05D1/0246, B25J5/007, G05D1/0272, G05D1/0225, B25J19/0091, G06N3/008, G06F19/3418, G05D1/0274, G05D1/0255, G05D1/0251, B25J19/023, G05D1/0234, G05D1/027, B25J19/06, Y10S901/47, Y10S901/01, B25J9/1697, G05D1/021|
|European Classification||G06F19/34C, G06F19/34A, B25J9/00D, G05D1/02E6V, B25J19/00S, B25J5/00W, G05D1/02E14M, B25J19/02B4, G05D1/02E14D, B25J19/06, G06F19/34L1, G06N3/00L3|
|Feb 8, 2011||CC||Certificate of correction|
|Mar 8, 2011||AS||Assignment|
Owner name: IROBOT CORPORATION, MASSACHUSETTS
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:VU, CLARA;CROSS, MATTHEW;CAMPBELL, TONY L.;SIGNING DATES FROM 20070417 TO 20091013;REEL/FRAME:025919/0943
|Nov 18, 2013||FPAY||Fee payment|
Year of fee payment: 4
|Aug 18, 2014||AS||Assignment|
Owner name: IROBOT CORPORATION, MASSACHUSETTS
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:ZIEGLER, ANDREW;JONES, ANDREW;VU, CLARA;AND OTHERS;SIGNING DATES FROM 20070417 TO 20140204;REEL/FRAME:033554/0558
|Oct 5, 2017||AS||Assignment|
Owner name: AVA ROBOTICS, INC., MASSACHUSETTS
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:IROBOT CORPORATION;REEL/FRAME:043794/0197
Effective date: 20170724